reactions string | labels string | performed_via_github_app string | milestone string | assignee string | html_url string | assignees string | timeline_url string | issue_field_values string | id string | repository_url string | url string | title string | issue_dependencies_summary string | created_at string | state_reason string | type string | sub_issues_summary string | author_association string | closed_by string | events_url string | locked string | closed_at string | pinned_comment string | body string | comments_url string | node_id string | active_lock_reason string | state string | comments string | pull_request string | updated_at string | labels_url string | number string | user string | draft string | is_paul_request bool |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8997/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8997 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8997/timeline | json:[] | value:6312903134 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8997 | value:Fix class_encode_column and align_labels_with_mapping on formatted datasets | value:2026-11-10T18:09:31Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8997/events | value:False | value:## Summary
Fixes #8996
`Dataset.class_encode_column` and `Dataset.align_labels_with_mapping` now work on formatted datasets. Before, with a `torch` format, `class_encode_column` on an integer column produced the class names `['tensor(1)', 'tensor(3)']` instead of `['1', '3']`, and with the `numpy`, `arrow` or `... | value:https://api.github.com/repos/huggingface/datasets/issues/8997/comments | value:PR_kwDODunzps8AAAABOLXYmA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8997", "html_url": "https://github.com/huggingface/datasets/pull/8997", "diff_url": "https://github.com/huggingface/datasets/pull/8997.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8997.patch", "merged_at": null} | value:2026-11-10T18:09:31Z | value:https://api.github.com/repos/huggingface/datasets/issues/8997/labels{/name} | value:8997 | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8996/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | value:https://github.com/huggingface/datasets/issues/8996 | json:[{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers",... | value:https://api.github.com/repos/huggingface/datasets/issues/8996/timeline | json:[] | value:6312870201 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8996 | value:class_encode_column and align_labels_with_mapping give wrong class names or fail on a formatted dataset | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-10T18:07:49Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8996/events | value:False | value:### Describe the bug
`Dataset.class_encode_column` and `Dataset.align_labels_with_mapping` run their internal `map` with the dataset's format, but the functions they map expect batches of python lists. On a dataset formatted with `torch`, `class_encode_column` on an integer column silently produces class names l... | value:https://api.github.com/repos/huggingface/datasets/issues/8996/comments | value:I_kwDODunzps8AAAABeEbBOQ | value:open | value:1 | value:2026-11-10T18:08:38Z | value:https://api.github.com/repos/huggingface/datasets/issues/8996/labels{/name} | value:8996 | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8995/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8995 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8995/timeline | json:[] | value:6308424421 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8995 | value:Raise instead of hanging when a num_proc worker dies in exports and downloads | value:2026-11-10T14:03:40Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8995/events | value:False | value:Fixes #8993.
With `num_proc > 1`, `Dataset.to_sql`, `to_csv` and `to_json` and the download/extract step of `load_dataset(..., num_proc=N)` never return if a pool worker dies abruptly, for example after an OOM kill, a `kill -9`, or a task that fails to unpickle in the worker. `Dataset.map` already raises in this... | value:https://api.github.com/repos/huggingface/datasets/issues/8995/comments | value:PR_kwDODunzps8AAAABOHwHrg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8995", "html_url": "https://github.com/huggingface/datasets/pull/8995", "diff_url": "https://github.com/huggingface/datasets/pull/8995.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8995.patch", "merged_at": null} | value:2026-11-10T14:03:40Z | value:https://api.github.com/repos/huggingface/datasets/issues/8995/labels{/name} | value:8995 | json:{"login": "AdrianPerez3", "id": 70107702, "node_id": "MDQ6VXNlcjcwMTA3NzAy", "avatar_url": "https://avatars.githubusercontent.com/u/70107702?v=4", "gravatar_id": "", "url": "https://api.github.com/users/AdrianPerez3", "html_url": "https://github.com/AdrianPerez3", "followers_url": "https://api.github.com/users/Adr... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8994/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8994 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8994/timeline | json:[] | value:6307247486 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8994 | value:Preserve zero affected-row counts in Dataset.to_sql | value:2026-11-10T12:55:30Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8994/events | value:False | value:`Dataset.to_sql` can report that rows were written when a custom pandas insertion method inserted none. For example, writing `[1, 2, 3, 4]` in batches of two with SQLite `INSERT OR IGNORE` returns `4` when the table already contains all four values. It also returns `4` instead of `2` when only the first batch is ... | value:https://api.github.com/repos/huggingface/datasets/issues/8994/comments | value:PR_kwDODunzps8AAAABOGz1Rg | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8994", "html_url": "https://github.com/huggingface/datasets/pull/8994", "diff_url": "https://github.com/huggingface/datasets/pull/8994.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8994.patch", "merged_at": null} | value:2026-11-10T13:48:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8994/labels{/name} | value:8994 | json:{"login": "anticipate218", "id": 185322635, "node_id": "U_kgDOCwvMiw", "avatar_url": "https://avatars.githubusercontent.com/u/185322635?v=4", "gravatar_id": "", "url": "https://api.github.com/users/anticipate218", "html_url": "https://github.com/anticipate218", "followers_url": "https://api.github.com/users/antici... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8993/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8993 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8993/timeline | json:[] | value:6303860921 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8993 | value:With `num_proc>1`, `to_sql`/`to_csv`/`to_json` and `load_dataset` downloads never return if a worker process dies | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-10T09:15:36Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8993/events | value:False | value:### Describe the bug
With `num_proc > 1`, two code paths send their tasks to a stdlib `multiprocessing.Pool` and wait on it without the dead-worker check that #5976 added to `iflatmap_unordered`:
- the writers behind `Dataset.to_sql`, `to_csv` and `to_json`, which iterate over `pool.imap(...)` (`io/sql.py`, `io... | value:https://api.github.com/repos/huggingface/datasets/issues/8993/comments | value:I_kwDODunzps8AAAABd71IuQ | value:open | value:2 | value:2026-11-10T10:13:42Z | value:https://api.github.com/repos/huggingface/datasets/issues/8993/labels{/name} | value:8993 | json:{"login": "AdrianPerez3", "id": 70107702, "node_id": "MDQ6VXNlcjcwMTA3NzAy", "avatar_url": "https://avatars.githubusercontent.com/u/70107702?v=4", "gravatar_id": "", "url": "https://api.github.com/users/AdrianPerez3", "html_url": "https://github.com/AdrianPerez3", "followers_url": "https://api.github.com/users/Adr... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8992/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8992 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8992/timeline | json:[] | value:6302413917 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8992 | value:Fix IterableDataset state_dict freezing when a sibling dataset is created | value:2026-11-10T07:35:29Z | value:CONTRIBUTOR | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8992/events | value:False | value:2026-11-10T17:23:43Z | value:Fixes #8991
Datasets derived from the same dataset (e.g. `ds.skip(n)` and `ds.take(n)`) share its ex_iterables. Creating or iterating one of them runs `_prepare_ex_iterable_for_iteration()`, which calls `_init_state_dict()` on the shared nodes and replaces their state dicts. The other dataset's running iterator ... | value:https://api.github.com/repos/huggingface/datasets/issues/8992/comments | value:PR_kwDODunzps8AAAABOC8P9w | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8992", "html_url": "https://github.com/huggingface/datasets/pull/8992", "diff_url": "https://github.com/huggingface/datasets/pull/8992.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8992.patch", "merged_at": "2026-11-10T17:23:43Z"} | value:2026-11-10T17:23:43Z | value:https://api.github.com/repos/huggingface/datasets/issues/8992/labels{/name} | value:8992 | json:{"login": "Dhruv-Sharma29", "id": 237243628, "node_id": "U_kgDODiQM7A", "avatar_url": "https://avatars.githubusercontent.com/u/237243628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Dhruv-Sharma29", "html_url": "https://github.com/Dhruv-Sharma29", "followers_url": "https://api.github.com/users/Dhr... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8991/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8991 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8991/timeline | json:[] | value:6302411182 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8991 | value:IterableDataset.state_dict() stops advancing when another dataset derived from the same source is created or iterated | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-10T07:35:16Z | value:completed | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8991/events | value:False | value:2026-11-10T17:23:45Z | value:### Describe the bug
If two datasets are derived from the same `IterableDataset` (for example `train_ds = ds.skip(n)` and `eval_ds = ds.take(n)`), creating or iterating one of them freezes the `state_dict()` of the other while it is being iterated. The examples it yields are still correct, but its checkpoint sto... | value:https://api.github.com/repos/huggingface/datasets/issues/8991/comments | value:I_kwDODunzps8AAAABd6cprg | value:closed | value:0 | value:2026-11-10T17:23:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8991/labels{/name} | value:8991 | json:{"login": "Dhruv-Sharma29", "id": 237243628, "node_id": "U_kgDODiQM7A", "avatar_url": "https://avatars.githubusercontent.com/u/237243628?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Dhruv-Sharma29", "html_url": "https://github.com/Dhruv-Sharma29", "followers_url": "https://api.github.com/users/Dhr... | true | ||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8990/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8990 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8990/timeline | json:[] | value:6302318501 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8990 | value:Fix formatting in README.md for clarity | value:2026-11-10T07:27:37Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8990/events | value:False | value:https://api.github.com/repos/huggingface/datasets/issues/8990/comments | value:PR_kwDODunzps8AAAABOC3cRg | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8990", "html_url": "https://github.com/huggingface/datasets/pull/8990", "diff_url": "https://github.com/huggingface/datasets/pull/8990.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8990.patch", "merged_at": null} | value:2026-11-10T07:32:54Z | value:https://api.github.com/repos/huggingface/datasets/issues/8990/labels{/name} | value:8990 | json:{"login": "iniya-27", "id": 154033315, "node_id": "U_kgDOCS5cow", "avatar_url": "https://avatars.githubusercontent.com/u/154033315?v=4", "gravatar_id": "", "url": "https://api.github.com/users/iniya-27", "html_url": "https://github.com/iniya-27", "followers_url": "https://api.github.com/users/iniya-27/followers", ... | value:False | true | ||||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8989/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8989 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8989/timeline | json:[] | value:6300796148 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8989 | value:Fix signature mismatch in FilteredExamplesIterable.shuffle_data_sources | value:2026-11-10T05:15:51Z | value:NONE | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | value:https://api.github.com/repos/huggingface/datasets/issues/8989/events | value:False | value:2026-11-10T05:16:12Z | value:Fixes #8988
### Description
Align `FilteredExamplesIterable.shuffle_data_sources` signature with `_BaseExamplesIterable` and sibling iterable classes (`MappedExamplesIterable`, `ExamplesIterable`, etc.).
Currently, `FilteredExamplesIterable.shuffle_data_sources` accepts `seed: Optional[int]` and passes it dire... | value:https://api.github.com/repos/huggingface/datasets/issues/8989/comments | value:PR_kwDODunzps8AAAABOBqrqw | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8989", "html_url": "https://github.com/huggingface/datasets/pull/8989", "diff_url": "https://github.com/huggingface/datasets/pull/8989.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8989.patch", "merged_at": null} | value:2026-11-10T05:16:12Z | value:https://api.github.com/repos/huggingface/datasets/issues/8989/labels{/name} | value:8989 | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8988/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8988 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8988/timeline | json:[] | value:6300794469 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8988 | value:Signature mismatch in FilteredExamplesIterable.shuffle_data_sources | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-10T05:15:42Z | value:not_planned | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | value:https://api.github.com/repos/huggingface/datasets/issues/8988/events | value:False | value:2026-11-10T05:16:15Z | value:### Describe the bug
In `datasets.iterable_dataset.FilteredExamplesIterable`, the `shuffle_data_sources` method has a signature mismatch compared to `_BaseExamplesIterable` and all other sibling iterable classes:
```python
class FilteredExamplesIterable(MappedExamplesIterable):
...
def shuffle_data_sour... | value:https://api.github.com/repos/huggingface/datasets/issues/8988/comments | value:I_kwDODunzps8AAAABd45-ZQ | value:closed | value:1 | value:2026-11-10T05:16:15Z | value:https://api.github.com/repos/huggingface/datasets/issues/8988/labels{/name} | value:8988 | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | true | ||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8987/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8987 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8987/timeline | json:[] | value:6299530302 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8987 | value:Validate weights in Split.subsplit | value:2026-11-10T03:19:35Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8987/events | value:False | value:Fixes #8986
Reject negative weights and weight lists without a positive value in `Split.subsplit(weighted=...)`. This replaces division by zero or invalid percent slices with a clear `ValueError`, while retaining support for a zero weight paired with a positive weight.
Added regression cases for invalid weights... | value:https://api.github.com/repos/huggingface/datasets/issues/8987/comments | value:PR_kwDODunzps8AAAABOAqqOw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8987", "html_url": "https://github.com/huggingface/datasets/pull/8987", "diff_url": "https://github.com/huggingface/datasets/pull/8987.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8987.patch", "merged_at": null} | value:2026-11-10T12:20:10Z | value:https://api.github.com/repos/huggingface/datasets/issues/8987/labels{/name} | value:8987 | json:{"login": "Gninjaa", "id": 149934057, "node_id": "U_kgDOCO_P6Q", "avatar_url": "https://avatars.githubusercontent.com/u/149934057?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Gninjaa", "html_url": "https://github.com/Gninjaa", "followers_url": "https://api.github.com/users/Gninjaa/followers", "fol... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8986/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "Gninjaa", "id": 149934057, "node_id": "U_kgDOCO_P6Q", "avatar_url": "https://avatars.githubusercontent.com/u/149934057?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Gninjaa", "html_url": "https://github.com/Gninjaa", "followers_url": "https://api.github.com/users/Gninjaa/followers", "fol... | value:https://github.com/huggingface/datasets/issues/8986 | json:[{"login": "Gninjaa", "id": 149934057, "node_id": "U_kgDOCO_P6Q", "avatar_url": "https://avatars.githubusercontent.com/u/149934057?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Gninjaa", "html_url": "https://github.com/Gninjaa", "followers_url": "https://api.github.com/users/Gninjaa/followers", "fo... | value:https://api.github.com/repos/huggingface/datasets/issues/8986/timeline | json:[] | value:6299518913 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8986 | value:Split.subsplit(weighted=...) should reject zero-total and negative weights | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-10T03:18:35Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8986/events | value:False | value:### Describe the bug
`Split.subsplit(weighted=...)` accepts negative weights and can produce invalid percent slices. When all weights are zero, it raises `ZeroDivisionError` while normalizing the weights.
### Steps to reproduce the bug
```python
from datasets import Split
Split.TRAIN.subsplit(weighted=[0, 0])... | value:https://api.github.com/repos/huggingface/datasets/issues/8986/comments | value:I_kwDODunzps8AAAABd3sHwQ | value:open | value:1 | value:2026-11-10T03:18:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8986/labels{/name} | value:8986 | json:{"login": "Gninjaa", "id": 149934057, "node_id": "U_kgDOCO_P6Q", "avatar_url": "https://avatars.githubusercontent.com/u/149934057?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Gninjaa", "html_url": "https://github.com/Gninjaa", "followers_url": "https://api.github.com/users/Gninjaa/followers", "fol... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8985/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8985 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8985/timeline | json:[] | value:6295287313 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8985 | value:set dev version | value:2026-11-09T21:41:28Z | value:MEMBER | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8985/events | value:False | value:2026-11-09T21:41:38Z | value:https://api.github.com/repos/huggingface/datasets/issues/8985/comments | value:PR_kwDODunzps8AAAABN9S_qg | value:closed | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8985", "html_url": "https://github.com/huggingface/datasets/pull/8985", "diff_url": "https://github.com/huggingface/datasets/pull/8985.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8985.patch", "merged_at": "2026-11-09T21:41:38Z"} | value:2026-11-09T21:41:41Z | value:https://api.github.com/repos/huggingface/datasets/issues/8985/labels{/name} | value:8985 | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:False | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8984/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8984 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8984/timeline | json:[] | value:6295213569 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8984 | value:release: 5.2.0 | value:2026-11-09T21:36:45Z | value:MEMBER | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8984/events | value:False | value:2026-11-09T21:38:04Z | value:https://api.github.com/repos/huggingface/datasets/issues/8984/comments | value:PR_kwDODunzps8AAAABN9PPOA | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8984", "html_url": "https://github.com/huggingface/datasets/pull/8984", "diff_url": "https://github.com/huggingface/datasets/pull/8984.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8984.patch", "merged_at": "2026-11-09T21:38:04Z"} | value:2026-11-09T21:38:33Z | value:https://api.github.com/repos/huggingface/datasets/issues/8984/labels{/name} | value:8984 | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:False | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8983/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8983 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8983/timeline | json:[] | value:6295158116 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8983 | value:Add SharedBlockCacheFileSystem | value:2026-11-09T21:33:02Z | value:MEMBER | value:https://api.github.com/repos/huggingface/datasets/issues/8983/events | value:False | value:could be useful for datasets that do tons of random access to files when streaming, like videos in robotics datasets in `lerobot` | value:https://api.github.com/repos/huggingface/datasets/issues/8983/comments | value:PR_kwDODunzps8AAAABN9MYJw | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8983", "html_url": "https://github.com/huggingface/datasets/pull/8983", "diff_url": "https://github.com/huggingface/datasets/pull/8983.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8983.patch", "merged_at": null} | value:2026-11-10T16:51:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8983/labels{/name} | value:8983 | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8982/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8982 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8982/timeline | json:[] | value:6290882341 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8982 | value:fix(image): keep bytes for relative archive labels | value:2026-11-09T17:11:01Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8982/events | value:False | value:## Summary
A local file with the same name as a WebDataset member can replace the embedded image. When an `Image` dictionary supplies both a relative archive label and bytes, the encoder can discard the bytes if the label happens to resolve to a file in the working directory.
Keep the relative label and the sup... | value:https://api.github.com/repos/huggingface/datasets/issues/8982/comments | value:PR_kwDODunzps8AAAABN5v5xw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8982", "html_url": "https://github.com/huggingface/datasets/pull/8982", "diff_url": "https://github.com/huggingface/datasets/pull/8982.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8982.patch", "merged_at": null} | value:2026-11-09T17:11:01Z | value:https://api.github.com/repos/huggingface/datasets/issues/8982/labels{/name} | value:8982 | json:{"login": "GoldArowana", "id": 27399349, "node_id": "MDQ6VXNlcjI3Mzk5MzQ5", "avatar_url": "https://avatars.githubusercontent.com/u/27399349?v=4", "gravatar_id": "", "url": "https://api.github.com/users/GoldArowana", "html_url": "https://github.com/GoldArowana", "followers_url": "https://api.github.com/users/GoldAr... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8981/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8981 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8981/timeline | json:[] | value:6274148965 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8981 | value:Fix IterableDataset.add_column with table formats | value:2026-11-08T18:26:32Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8981/events | value:False | value:Fixes #8980.
Adding a column to an Arrow-, pandas-, or Polars-formatted `IterableDataset` currently raises `TypeError` during iteration because the internal callback returns a dictionary. Perform this operation in batched Arrow format, append the new column without discarding the original columns, and restore th... | value:https://api.github.com/repos/huggingface/datasets/issues/8981/comments | value:PR_kwDODunzps8AAAABNsYR6g | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8981", "html_url": "https://github.com/huggingface/datasets/pull/8981", "diff_url": "https://github.com/huggingface/datasets/pull/8981.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8981.patch", "merged_at": null} | value:2026-11-08T18:26:32Z | value:https://api.github.com/repos/huggingface/datasets/issues/8981/labels{/name} | value:8981 | json:{"login": "FOWEPJF255", "id": 270070255, "node_id": "U_kgDOEBjx7w", "avatar_url": "https://avatars.githubusercontent.com/u/270070255?v=4", "gravatar_id": "", "url": "https://api.github.com/users/FOWEPJF255", "html_url": "https://github.com/FOWEPJF255", "followers_url": "https://api.github.com/users/FOWEPJF255/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8980/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "FOWEPJF255", "id": 270070255, "node_id": "U_kgDOEBjx7w", "avatar_url": "https://avatars.githubusercontent.com/u/270070255?v=4", "gravatar_id": "", "url": "https://api.github.com/users/FOWEPJF255", "html_url": "https://github.com/FOWEPJF255", "followers_url": "https://api.github.com/users/FOWEPJF255/foll... | value:https://github.com/huggingface/datasets/issues/8980 | json:[{"login": "FOWEPJF255", "id": 270070255, "node_id": "U_kgDOEBjx7w", "avatar_url": "https://avatars.githubusercontent.com/u/270070255?v=4", "gravatar_id": "", "url": "https://api.github.com/users/FOWEPJF255", "html_url": "https://github.com/FOWEPJF255", "followers_url": "https://api.github.com/users/FOWEPJF255/fol... | value:https://api.github.com/repos/huggingface/datasets/issues/8980/timeline | json:[] | value:6274107647 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8980 | value:IterableDataset.add_column fails with arrow, pandas and polars formats | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-08T18:21:59Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8980/events | value:False | value:### Describe the bug
`IterableDataset.add_column()` fails during iteration after `with_format("arrow")`, `with_format("pandas")`, or `with_format("polars")`. Adding the column before selecting the format works. This prevents a streaming data preparation pipeline from adding metadata columns after it has selected... | value:https://api.github.com/repos/huggingface/datasets/issues/8980/comments | value:I_kwDODunzps8AAAABdfdI_w | value:open | value:1 | value:2026-11-08T18:22:19Z | value:https://api.github.com/repos/huggingface/datasets/issues/8980/labels{/name} | value:8980 | json:{"login": "FOWEPJF255", "id": 270070255, "node_id": "U_kgDOEBjx7w", "avatar_url": "https://avatars.githubusercontent.com/u/270070255?v=4", "gravatar_id": "", "url": "https://api.github.com/users/FOWEPJF255", "html_url": "https://github.com/FOWEPJF255", "followers_url": "https://api.github.com/users/FOWEPJF255/foll... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8979/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8979 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8979/timeline | json:[] | value:6269880534 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8979 | value:Preserve feature conversion when combining typed iterable sources | value:2026-11-08T10:13:20Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8979/events | value:False | value:Combining iterable datasets can silently skip feature conversion or omit columns depending on the first source. For example, concatenating a source processed with `.map(..., features=features)` with a raw generator declaring the same `ClassLabel` produces integer labels from the first source and strings from the ... | value:https://api.github.com/repos/huggingface/datasets/issues/8979/comments | value:PR_kwDODunzps8AAAABNpGJrA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8979", "html_url": "https://github.com/huggingface/datasets/pull/8979", "diff_url": "https://github.com/huggingface/datasets/pull/8979.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8979.patch", "merged_at": null} | value:2026-11-08T10:13:29Z | value:https://api.github.com/repos/huggingface/datasets/issues/8979/labels{/name} | value:8979 | json:{"login": "remote-controlled-man", "id": 134844631, "node_id": "U_kgDOCAmQ1w", "avatar_url": "https://avatars.githubusercontent.com/u/134844631?v=4", "gravatar_id": "", "url": "https://api.github.com/users/remote-controlled-man", "html_url": "https://github.com/remote-controlled-man", "followers_url": "https://api... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8978/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8978 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8978/timeline | json:[] | value:6269262100 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8978 | value:Fix loaded PIL encoding after local source removal | value:2026-11-08T08:48:25Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8978/events | value:False | value:A fully loaded PIL image remains usable after its local source file is removed. `encode_pil_image` currently stores the retained filename, so `Dataset.from_dict` and a `Dataset.map` callback can produce a dataset that raises `FileNotFoundError` on access or `save_to_disk` despite having received usable image pixe... | value:https://api.github.com/repos/huggingface/datasets/issues/8978/comments | value:PR_kwDODunzps8AAAABNon9SA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8978", "html_url": "https://github.com/huggingface/datasets/pull/8978", "diff_url": "https://github.com/huggingface/datasets/pull/8978.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8978.patch", "merged_at": null} | value:2026-11-08T08:48:25Z | value:https://api.github.com/repos/huggingface/datasets/issues/8978/labels{/name} | value:8978 | json:{"login": "GoldArowana", "id": 27399349, "node_id": "MDQ6VXNlcjI3Mzk5MzQ5", "avatar_url": "https://avatars.githubusercontent.com/u/27399349?v=4", "gravatar_id": "", "url": "https://api.github.com/users/GoldArowana", "html_url": "https://github.com/GoldArowana", "followers_url": "https://api.github.com/users/GoldAr... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8977/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8977 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8977/timeline | json:[] | value:6266654536 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8977 | value:Fix IterableDataset.rename_columns with arrow, pandas and polars formats | value:2026-11-08T02:17:41Z | value:NONE | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8977/events | value:False | value:2026-11-10T17:24:13Z | value:`IterableDataset.rename_column` / `rename_columns` crash during iteration when the dataset has a table format (`arrow`, `pandas`, `polars`).
```python
from datasets import IterableDataset
ds = IterableDataset.from_dict({"a": [1, 2, 3], "b": ["x", "y", "z"]})
list(ds.with_format("arrow").rename_column("a",... | value:https://api.github.com/repos/huggingface/datasets/issues/8977/comments | value:PR_kwDODunzps8AAAABNmnIdQ | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8977", "html_url": "https://github.com/huggingface/datasets/pull/8977", "diff_url": "https://github.com/huggingface/datasets/pull/8977.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8977.patch", "merged_at": "2026-11-10T17:24:13Z"} | value:2026-11-10T17:24:13Z | value:https://api.github.com/repos/huggingface/datasets/issues/8977/labels{/name} | value:8977 | json:{"login": "assass1nlike", "id": 326038984, "node_id": "U_kgDOE271yA", "avatar_url": "https://avatars.githubusercontent.com/u/326038984?v=4", "gravatar_id": "", "url": "https://api.github.com/users/assass1nlike", "html_url": "https://github.com/assass1nlike", "followers_url": "https://api.github.com/users/assass1nl... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8976/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8976 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8976/timeline | json:[] | value:6266442521 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8976 | value:Fix mixed-type JSON fields with slashes in their names | value:2026-11-08T01:49:10Z | value:CONTRIBUTOR | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8976/events | value:False | value:2026-11-09T17:28:50Z | value:Fixes #8898
cc @lhoestq | value:https://api.github.com/repos/huggingface/datasets/issues/8976/comments | value:PR_kwDODunzps8AAAABNmcdng | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8976", "html_url": "https://github.com/huggingface/datasets/pull/8976", "diff_url": "https://github.com/huggingface/datasets/pull/8976.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8976.patch", "merged_at": "2026-11-09T17:28:50Z"} | value:2026-11-09T17:28:50Z | value:https://api.github.com/repos/huggingface/datasets/issues/8976/labels{/name} | value:8976 | json:{"login": "its-amann", "id": 163747435, "node_id": "U_kgDOCcKWaw", "avatar_url": "https://avatars.githubusercontent.com/u/163747435?v=4", "gravatar_id": "", "url": "https://api.github.com/users/its-amann", "html_url": "https://github.com/its-amann", "followers_url": "https://api.github.com/users/its-amann/follower... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8975/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8975 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8975/timeline | json:[] | value:6266254101 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8975 | value:Fix Parquet compression disabled for all columns in structs containing media | value:2026-11-08T01:21:44Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8975/events | value:False | value:Fixes #8973
\get_parquet_column_options\ applied a single top-level \
equire_storage_embed\
result to every leaf under a struct field, disabling compression and dictionary
encoding for non-media columns whenever a media feature was present. Compute the
embed flag per leaf based on its own feature.
This pull req... | value:https://api.github.com/repos/huggingface/datasets/issues/8975/comments | value:PR_kwDODunzps8AAAABNmTHlw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8975", "html_url": "https://github.com/huggingface/datasets/pull/8975", "diff_url": "https://github.com/huggingface/datasets/pull/8975.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8975.patch", "merged_at": null} | value:2026-11-08T01:21:44Z | value:https://api.github.com/repos/huggingface/datasets/issues/8975/labels{/name} | value:8975 | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8974/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8974 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8974/timeline | json:[] | value:6266018236 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8974 | value:Compress nested non-media columns in Parquet writer alongside media features | value:2026-11-08T00:50:48Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8974/events | value:False | value:Fixes #8973
### Context & Problem
In #8674, Parquet writer options (`compression`, `use_dictionary`, `column_encoding`) were keyed by leaf path rather than top-level column names so nested struct/list fields would be compressed.
However, `get_parquet_column_options` still checked `require_storage_embed` only a... | value:https://api.github.com/repos/huggingface/datasets/issues/8974/comments | value:PR_kwDODunzps8AAAABNmHV5w | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8974", "html_url": "https://github.com/huggingface/datasets/pull/8974", "diff_url": "https://github.com/huggingface/datasets/pull/8974.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8974.patch", "merged_at": null} | value:2026-11-08T00:50:48Z | value:https://api.github.com/repos/huggingface/datasets/issues/8974/labels{/name} | value:8974 | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8973/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | value:https://github.com/huggingface/datasets/issues/8973 | json:[{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/fol... | value:https://api.github.com/repos/huggingface/datasets/issues/8973/timeline | json:[] | value:6266008854 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8973 | value:Parquet writer disables compression for non-media columns in structs containing an Image or Audio feature | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-08T00:49:28Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8973/events | value:False | value:### Describe the bug
When a dataset contains a struct or dictionary column that includes an embedded storage feature (such as `Image`, `Audio`, `Video`, etc.) alongside non-media columns (such as text strings, integers, floats, or vectors), saving the dataset to Parquet via `to_parquet()` or `push_to_hub()` disa... | value:https://api.github.com/repos/huggingface/datasets/issues/8973/comments | value:I_kwDODunzps8AAAABdXu1Fg | value:open | value:1 | value:2026-11-08T00:49:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8973/labels{/name} | value:8973 | json:{"login": "davrens101", "id": 166565828, "node_id": "U_kgDOCe2XxA", "avatar_url": "https://avatars.githubusercontent.com/u/166565828?v=4", "gravatar_id": "", "url": "https://api.github.com/users/davrens101", "html_url": "https://github.com/davrens101", "followers_url": "https://api.github.com/users/davrens101/foll... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8972/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8972 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8972/timeline | json:[] | value:6265890421 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8972 | value:Fix the eval loader for string sample ids | value:2026-11-08T00:34:34Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8972/events | value:False | value:The `eval` loader can't read Inspect logs whose sample ids are strings, which many tasks use (SWE-bench, GPQA, ...). Inspect names sample files `{sample_id}_epoch_{epoch}.json`, and `_sort_samples_key` calls `int()` on the id:
```python
from datasets import load_dataset
load_dataset("eval", data_files="swe_benc... | value:https://api.github.com/repos/huggingface/datasets/issues/8972/comments | value:PR_kwDODunzps8AAAABNmA7Ag | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8972", "html_url": "https://github.com/huggingface/datasets/pull/8972", "diff_url": "https://github.com/huggingface/datasets/pull/8972.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8972.patch", "merged_at": null} | value:2026-11-08T11:13:13Z | value:https://api.github.com/repos/huggingface/datasets/issues/8972/labels{/name} | value:8972 | json:{"login": "aidxhxr", "id": 93529551, "node_id": "U_kgDOBZMlzw", "avatar_url": "https://avatars.githubusercontent.com/u/93529551?v=4", "gravatar_id": "", "url": "https://api.github.com/users/aidxhxr", "html_url": "https://github.com/aidxhxr", "followers_url": "https://api.github.com/users/aidxhxr/followers", "follo... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8971/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8971 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8971/timeline | json:[] | value:6265849887 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8971 | value:docs: fix unecessary and atleast typos in tests | value:2026-11-08T00:29:00Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8971/events | value:False | value:Correct `unecessary` to `unnecessary` in test comments and `atleast` to `at least` in an assertion message. | value:https://api.github.com/repos/huggingface/datasets/issues/8971/comments | value:PR_kwDODunzps8AAAABNl-6Pg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8971", "html_url": "https://github.com/huggingface/datasets/pull/8971", "diff_url": "https://github.com/huggingface/datasets/pull/8971.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8971.patch", "merged_at": null} | value:2026-11-08T00:29:00Z | value:https://api.github.com/repos/huggingface/datasets/issues/8971/labels{/name} | value:8971 | json:{"login": "ninglinLiu", "id": 224592858, "node_id": "U_kgDODWMD2g", "avatar_url": "https://avatars.githubusercontent.com/u/224592858?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ninglinLiu", "html_url": "https://github.com/ninglinLiu", "followers_url": "https://api.github.com/users/ninglinLiu/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8970/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8970 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8970/timeline | json:[] | value:6261052761 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8970 | value:Document how to load Parquet files from the Dataset Viewer | value:2026-11-07T15:48:21Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8970/events | value:False | value:Closes #3220.
The Hub loading guide now shows where to find the Dataset Viewer’s **Auto-converted to Parquet** link and how to load one split from the conversion branch. The example uses the same CSV dataset already introduced in the guide and links to the Hub documentation for partial conversions and API access... | value:https://api.github.com/repos/huggingface/datasets/issues/8970/comments | value:PR_kwDODunzps8AAAABNiOk8g | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8970", "html_url": "https://github.com/huggingface/datasets/pull/8970", "diff_url": "https://github.com/huggingface/datasets/pull/8970.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8970.patch", "merged_at": null} | value:2026-11-07T15:48:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8970/labels{/name} | value:8970 | json:{"login": "Lizidong01", "id": 61378198, "node_id": "MDQ6VXNlcjYxMzc4MTk4", "avatar_url": "https://avatars.githubusercontent.com/u/61378198?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Lizidong01", "html_url": "https://github.com/Lizidong01", "followers_url": "https://api.github.com/users/Lizidong0... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8969/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8969 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8969/timeline | json:[] | value:6258743071 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8969 | value:Respect parent struct nulls when embedding media storage | value:2026-11-07T11:55:27Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8969/events | value:False | value:Saving a dataset containing a null struct can try to read media paths hidden beneath that null. Arrow child arrays may retain values under a null parent; those values are not part of the logical dataset. If a hidden path does not exist, `save_to_disk` raises `FileNotFoundError` even though the corresponding datas... | value:https://api.github.com/repos/huggingface/datasets/issues/8969/comments | value:PR_kwDODunzps8AAAABNgZ0KQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8969", "html_url": "https://github.com/huggingface/datasets/pull/8969", "diff_url": "https://github.com/huggingface/datasets/pull/8969.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8969.patch", "merged_at": null} | value:2026-11-07T11:55:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8969/labels{/name} | value:8969 | json:{"login": "remote-controlled-man", "id": 134844631, "node_id": "U_kgDOCAmQ1w", "avatar_url": "https://avatars.githubusercontent.com/u/134844631?v=4", "gravatar_id": "", "url": "https://api.github.com/users/remote-controlled-man", "html_url": "https://github.com/remote-controlled-man", "followers_url": "https://api... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8968/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8968 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8968/timeline | json:[] | value:6258644043 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8968 | value:Fix map with num_proc failing on a cache hit when suffix_template has a custom rank format | value:2026-11-07T11:44:12Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8968/events | value:False | value:## Summary
Fixes #8967
`Dataset.map(num_proc=...)` with a `suffix_template` whose rank field is not exactly `{rank:05d}` (for example `_{rank:03d}_of_{num_proc:03d}`) worked on the first run, then raised `ValueError: Unknown format code 'd' for object of type 'str'` on every rerun that loaded all shards from th... | value:https://api.github.com/repos/huggingface/datasets/issues/8968/comments | value:PR_kwDODunzps8AAAABNgU1jw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8968", "html_url": "https://github.com/huggingface/datasets/pull/8968", "diff_url": "https://github.com/huggingface/datasets/pull/8968.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8968.patch", "merged_at": null} | value:2026-11-07T11:44:12Z | value:https://api.github.com/repos/huggingface/datasets/issues/8968/labels{/name} | value:8968 | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8967/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | value:https://github.com/huggingface/datasets/issues/8967 | json:[{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers",... | value:https://api.github.com/repos/huggingface/datasets/issues/8967/timeline | json:[] | value:6258629802 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8967 | value:map(num_proc=...) crashes on a cache hit when suffix_template uses a rank format other than {rank:05d} | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-07T11:42:34Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8967/events | value:False | value:### Describe the bug
`Dataset.map` with `num_proc` and a custom `suffix_template` works the first time, but every rerun that loads all shards from the cache raises `ValueError: Unknown format code 'd' for object of type 'str'`. Only the default `{rank:05d}` works. `Dataset.filter` has the same `suffix_template` ... | value:https://api.github.com/repos/huggingface/datasets/issues/8967/comments | value:I_kwDODunzps8AAAABdQscqg | value:open | value:1 | value:2026-11-07T11:42:56Z | value:https://api.github.com/repos/huggingface/datasets/issues/8967/labels{/name} | value:8967 | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8966/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8966 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8966/timeline | json:[] | value:6252884195 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8966 | value:Fix batch(by_column=) merging groups that are separated by null values | value:2026-11-07T00:02:33Z | value:NONE | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8966/events | value:False | value:2026-11-10T17:24:28Z | value:## Summary
Fixes #8965
`Dataset.batch(by_column=...)` and `IterableDataset.batch(by_column=...)` treated a null key as equal to its neighbours, so different groups were merged into one batch. Keys `[1, None, 2]` gave a single batch. A key column with only nulls raised `ArrowNotImplementedError`. Now a null is i... | value:https://api.github.com/repos/huggingface/datasets/issues/8966/comments | value:PR_kwDODunzps8AAAABNbzIaw | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8966", "html_url": "https://github.com/huggingface/datasets/pull/8966", "diff_url": "https://github.com/huggingface/datasets/pull/8966.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8966.patch", "merged_at": "2026-11-10T17:24:28Z"} | value:2026-11-10T17:24:28Z | value:https://api.github.com/repos/huggingface/datasets/issues/8966/labels{/name} | value:8966 | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8965/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | value:https://github.com/huggingface/datasets/issues/8965 | json:[{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers",... | value:https://api.github.com/repos/huggingface/datasets/issues/8965/timeline | json:[] | value:6252844094 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8965 | value:batch(by_column=...) merges different groups when the key column contains nulls | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-06T23:58:39Z | value:completed | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8965/events | value:False | value:2026-11-10T17:24:29Z | value:### Describe the bug
`Dataset.batch(by_column=...)` and `IterableDataset.batch(by_column=...)` treat a null key as equal to its neighbours, so rows from different groups end up in the same batch. A key column that contains only nulls raises instead.
### Steps to reproduce the bug
```python
from datasets import... | value:https://api.github.com/repos/huggingface/datasets/issues/8965/comments | value:I_kwDODunzps8AAAABdLLUPg | value:closed | value:1 | value:2026-11-10T17:24:29Z | value:https://api.github.com/repos/huggingface/datasets/issues/8965/labels{/name} | value:8965 | json:{"login": "Shizoqua", "id": 136805224, "node_id": "U_kgDOCCd7aA", "avatar_url": "https://avatars.githubusercontent.com/u/136805224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Shizoqua", "html_url": "https://github.com/Shizoqua", "followers_url": "https://api.github.com/users/Shizoqua/followers", ... | true | |||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8964/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8964 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8964/timeline | json:[] | value:6247326162 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8964 | value:Fix audio array tensor conversion with torch formatting | value:2026-11-06T17:18:36Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8964/events | value:False | value:## Description
Fixes #8755 — Audio arrays remain NumPy arrays when using PyTorch formatting.
### Changes made
- Updated the TorchCodec audio decoder to support returning PyTorch tensors.
- Updated `TorchFormatter` to convert Hugging Face audio decoder output to PyTorch-compatible output.
- Preserved existing Nu... | value:https://api.github.com/repos/huggingface/datasets/issues/8964/comments | value:PR_kwDODunzps8AAAABNXSGnw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8964", "html_url": "https://github.com/huggingface/datasets/pull/8964", "diff_url": "https://github.com/huggingface/datasets/pull/8964.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8964.patch", "merged_at": null} | value:2026-11-07T17:23:29Z | value:https://api.github.com/repos/huggingface/datasets/issues/8964/labels{/name} | value:8964 | json:{"login": "josephabishak88", "id": 249288588, "node_id": "U_kgDODtvXjA", "avatar_url": "https://avatars.githubusercontent.com/u/249288588?v=4", "gravatar_id": "", "url": "https://api.github.com/users/josephabishak88", "html_url": "https://github.com/josephabishak88", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8963/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8963 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8963/timeline | json:[] | value:6223147260 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8963 | value:Fix Spark working_dir, tenbin infinite loop, HF auth host check, missing aiohttp and Harbor errors | value:2026-11-05T12:26:07Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8963/events | value:False | value:I read through the repo and found a handful of small bugs. I reproduced each one before writing a fix. They're independent, so I can split this into separate PRs if that's easier to review.
### `http(s)://` data files fail on a clean install
fsspec's `HTTPFileSystem` needs `aiohttp`, but `aiohttp` is only in th... | value:https://api.github.com/repos/huggingface/datasets/issues/8963/comments | value:PR_kwDODunzps8AAAABND8Yng | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8963", "html_url": "https://github.com/huggingface/datasets/pull/8963", "diff_url": "https://github.com/huggingface/datasets/pull/8963.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8963.patch", "merged_at": null} | value:2026-11-05T12:26:07Z | value:https://api.github.com/repos/huggingface/datasets/issues/8963/labels{/name} | value:8963 | json:{"login": "Marouane-elfadli", "id": 228027721, "node_id": "U_kgDODZdtSQ", "avatar_url": "https://avatars.githubusercontent.com/u/228027721?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Marouane-elfadli", "html_url": "https://github.com/Marouane-elfadli", "followers_url": "https://api.github.com/use... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8962/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8962 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8962/timeline | json:[] | value:6218188671 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8962 | value:Fix : 3D format detection when texture files outnumber models | value:2026-11-05T06:39:31Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8962/events | value:False | value:Closes #8956
### Summary
- Prioritize 3D dataset detection when a repository contains mesh files alongside image textures.
- Prevent image textures from causing a 3D dataset to be incorrectly detected as an `imagefolder`.
- Preserve the existing format-selection behavior for other formats.
### Tests
- `python -... | value:https://api.github.com/repos/huggingface/datasets/issues/8962/comments | value:PR_kwDODunzps8AAAABM_8D8g | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8962", "html_url": "https://github.com/huggingface/datasets/pull/8962", "diff_url": "https://github.com/huggingface/datasets/pull/8962.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8962.patch", "merged_at": null} | value:2026-11-05T06:39:31Z | value:https://api.github.com/repos/huggingface/datasets/issues/8962/labels{/name} | value:8962 | json:{"login": "bhumika-231206", "id": 351443745, "node_id": "U_kgDOFPKbIQ", "avatar_url": "https://avatars.githubusercontent.com/u/351443745?v=4", "gravatar_id": "", "url": "https://api.github.com/users/bhumika-231206", "html_url": "https://github.com/bhumika-231206", "followers_url": "https://api.github.com/users/bhu... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8961/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8961 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8961/timeline | json:[] | value:6212024072 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8961 | value:Fix appending declared uint64 values beyond int64 range | value:2026-11-04T21:48:27Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8961/events | value:False | value:`Dataset.add_item` infers an Arrow table before applying the existing dataset features. Appending a Python integer above `2**63 - 1` therefore raises `OverflowError` even when its column is declared as `Value("uint64")`.
Retry the overflowing column with its existing scalar uint64 type. Other columns retain thei... | value:https://api.github.com/repos/huggingface/datasets/issues/8961/comments | value:PR_kwDODunzps8AAAABM7A3nA | value:open | value:3 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8961", "html_url": "https://github.com/huggingface/datasets/pull/8961", "diff_url": "https://github.com/huggingface/datasets/pull/8961.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8961.patch", "merged_at": null} | value:2026-11-06T10:00:51Z | value:https://api.github.com/repos/huggingface/datasets/issues/8961/labels{/name} | value:8961 | json:{"login": "GoldArowana", "id": 27399349, "node_id": "MDQ6VXNlcjI3Mzk5MzQ5", "avatar_url": "https://avatars.githubusercontent.com/u/27399349?v=4", "gravatar_id": "", "url": "https://api.github.com/users/GoldArowana", "html_url": "https://github.com/GoldArowana", "followers_url": "https://api.github.com/users/GoldAr... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8960/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8960 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8960/timeline | json:[] | value:6206755398 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8960 | value:Fix batch(by_column=) default accumulation causing full downloads | value:2026-11-04T16:20:14Z | value:MEMBER | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8960/events | value:False | value:2026-11-04T16:20:33Z | value:It was using the default of `batch_size` which is `None`, which means full downloads
By default it now uses max batch size (1000) | value:https://api.github.com/repos/huggingface/datasets/issues/8960/comments | value:PR_kwDODunzps8AAAABM2vxEg | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8960", "html_url": "https://github.com/huggingface/datasets/pull/8960", "diff_url": "https://github.com/huggingface/datasets/pull/8960.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8960.patch", "merged_at": "2026-11-04T16:20:33Z"} | value:2026-11-04T16:28:11Z | value:https://api.github.com/repos/huggingface/datasets/issues/8960/labels{/name} | value:8960 | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8959/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8959 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8959/timeline | json:[] | value:6206566552 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8959 | value:Fix CSV loading with explicit timestamp features | value:2026-11-04T16:08:14Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8959/events | value:False | value:CSV loading with an explicit `Value("timestamp[ns]")` feature fails in the default pandas C parser: the feature's NumPy datetime dtype is unsupported by `read_csv(dtype=...)`.
Read timestamp columns as objects and let the existing Arrow schema conversion parse them. This preserves the declared timestamp type, na... | value:https://api.github.com/repos/huggingface/datasets/issues/8959/comments | value:PR_kwDODunzps8AAAABM2mErg | value:open | value:2 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8959", "html_url": "https://github.com/huggingface/datasets/pull/8959", "diff_url": "https://github.com/huggingface/datasets/pull/8959.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8959.patch", "merged_at": null} | value:2026-11-05T14:32:20Z | value:https://api.github.com/repos/huggingface/datasets/issues/8959/labels{/name} | value:8959 | json:{"login": "feiiiiii5", "id": 204683769, "node_id": "U_kgDODDM5-Q", "avatar_url": "https://avatars.githubusercontent.com/u/204683769?v=4", "gravatar_id": "", "url": "https://api.github.com/users/feiiiiii5", "html_url": "https://github.com/feiiiiii5", "followers_url": "https://api.github.com/users/feiiiiii5/follower... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8958/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8958 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8958/timeline | json:[] | value:6204466923 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8958 | value:fix(mesh): accept .gltf in the Mesh feature and the meshfolder builder | value:2026-11-04T14:13:03Z | value:CONTRIBUTOR | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8958/events | value:False | value:2026-11-04T17:56:43Z | value:`Mesh` (and the `meshfolder` builder behind it) only accepts `.glb`, `.ply` and `.stl`, although `trimesh` — which does the decoding — also reads glTF. So a folder of `.gltf` models produces no `Mesh` column at all, and `Mesh().decode_example({"path": "model.gltf", ...})` raises.
Measured on `main` (62e0b421), b... | value:https://api.github.com/repos/huggingface/datasets/issues/8958/comments | value:PR_kwDODunzps8AAAABM040tw | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8958", "html_url": "https://github.com/huggingface/datasets/pull/8958", "diff_url": "https://github.com/huggingface/datasets/pull/8958.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8958.patch", "merged_at": "2026-11-04T17:56:43Z"} | value:2026-11-04T17:56:43Z | value:https://api.github.com/repos/huggingface/datasets/issues/8958/labels{/name} | value:8958 | json:{"login": "li-lizhe", "id": 82501526, "node_id": "MDQ6VXNlcjgyNTAxNTI2", "avatar_url": "https://avatars.githubusercontent.com/u/82501526?v=4", "gravatar_id": "", "url": "https://api.github.com/users/li-lizhe", "html_url": "https://github.com/li-lizhe", "followers_url": "https://api.github.com/users/li-lizhe/follow... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8957/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8957 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8957/timeline | json:[] | value:6204284026 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8957 | value:more distributed docs (hsdp) | value:2026-11-04T14:02:38Z | value:MEMBER | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8957/events | value:False | value:2026-11-05T11:28:05Z | value:cc @3outeille lmk if it sounds good to you
if `split_dataset_by_node` sounds confusing in this setting we can also imagine a different function dedicated to meshes | value:https://api.github.com/repos/huggingface/datasets/issues/8957/comments | value:PR_kwDODunzps8AAAABM0vWwQ | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8957", "html_url": "https://github.com/huggingface/datasets/pull/8957", "diff_url": "https://github.com/huggingface/datasets/pull/8957.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8957.patch", "merged_at": "2026-11-05T11:28:05Z"} | value:2026-11-05T11:28:08Z | value:https://api.github.com/repos/huggingface/datasets/issues/8957/labels{/name} | value:8957 | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8956/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8956 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8956/timeline | json:[] | value:6203429316 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8956 | value:[3d datasets]: repos with gltf textures get detected as images | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-04T13:12:16Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8956/events | value:False | value:[cfahlgren1/3d-model-samples](https://huggingface.co/datasets/cfahlgren1/3d-model-samples/tree/main) has 6 `.glb` files and one `.gltf` model, but the viewer shows it as an `Image` dataset with 15 rows of FlightHelmet textures and no 3D models.
The format comes from the most common file extension, and a `.gltf` ... | value:https://api.github.com/repos/huggingface/datasets/issues/8956/comments | value:I_kwDODunzps8AAAABccDRxA | value:open | value:3 | value:2026-11-05T04:42:54Z | value:https://api.github.com/repos/huggingface/datasets/issues/8956/labels{/name} | value:8956 | json:{"login": "cfahlgren1", "id": 13546028, "node_id": "MDQ6VXNlcjEzNTQ2MDI4", "avatar_url": "https://avatars.githubusercontent.com/u/13546028?v=4", "gravatar_id": "", "url": "https://api.github.com/users/cfahlgren1", "html_url": "https://github.com/cfahlgren1", "followers_url": "https://api.github.com/users/cfahlgren... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8955/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8955 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8955/timeline | json:[] | value:6195755633 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8955 | value:Validate range index bounds without scanning every index | value:2026-11-04T03:33:00Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8955/events | value:False | value:`ds.with_format("arrow")[range(n)]` can use an Arrow slice, but first `_check_valid_index_key` walks the entire range twice to compute its minimum and maximum. A range is monotonic, so its first and last elements give the same bounds for either step direction. Check the larger endpoint first as before, preserving... | value:https://api.github.com/repos/huggingface/datasets/issues/8955/comments | value:PR_kwDODunzps8AAAABMt4eQg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8955", "html_url": "https://github.com/huggingface/datasets/pull/8955", "diff_url": "https://github.com/huggingface/datasets/pull/8955.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8955.patch", "merged_at": null} | value:2026-11-04T03:33:00Z | value:https://api.github.com/repos/huggingface/datasets/issues/8955/labels{/name} | value:8955 | json:{"login": "justavibedev", "id": 213012880, "node_id": "U_kgDODLJRkA", "avatar_url": "https://avatars.githubusercontent.com/u/213012880?v=4", "gravatar_id": "", "url": "https://api.github.com/users/justavibedev", "html_url": "https://github.com/justavibedev", "followers_url": "https://api.github.com/users/justavibe... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8954/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8954 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8954/timeline | json:[] | value:6194424635 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8954 | value:Fix pickling round-trips for empty ConcatenationTable slices | value:2026-11-04T01:29:41Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8954/events | value:False | value:`ConcatenationTable.slice(0, 0)` returns the expected empty table, but a pickle round trip raises `ArrowInvalid: Must pass at least one table`.
Empty slices retain no row blocks. Rebuild those results directly from the saved schema, preserving column types, nullability and metadata. Nonempty block reconstruction... | value:https://api.github.com/repos/huggingface/datasets/issues/8954/comments | value:PR_kwDODunzps8AAAABMs0sZQ | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8954", "html_url": "https://github.com/huggingface/datasets/pull/8954", "diff_url": "https://github.com/huggingface/datasets/pull/8954.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8954.patch", "merged_at": null} | value:2026-11-05T13:39:08Z | value:https://api.github.com/repos/huggingface/datasets/issues/8954/labels{/name} | value:8954 | json:{"login": "ElMagAmine", "id": 36851541, "node_id": "MDQ6VXNlcjM2ODUxNTQx", "avatar_url": "https://avatars.githubusercontent.com/u/36851541?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ElMagAmine", "html_url": "https://github.com/ElMagAmine", "followers_url": "https://api.github.com/users/ElMagAmin... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8953/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8953 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8953/timeline | json:[] | value:6194313975 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8953 | value:Keep the loaded state in IterableDataset.state_dict() until iteration resumes | value:2026-11-04T01:19:34Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8953/events | value:False | value:`IterableDataset.load_state_dict()` only stores the state; it is applied when the next iteration starts. Until the first example is read, `state_dict()` keeps returning the dataset's previous state (for a dataset created after a restart, the start of the stream). A checkpoint saved in that window loses the restor... | value:https://api.github.com/repos/huggingface/datasets/issues/8953/comments | value:PR_kwDODunzps8AAAABMsvDww | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8953", "html_url": "https://github.com/huggingface/datasets/pull/8953", "diff_url": "https://github.com/huggingface/datasets/pull/8953.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8953.patch", "merged_at": null} | value:2026-11-05T13:41:34Z | value:https://api.github.com/repos/huggingface/datasets/issues/8953/labels{/name} | value:8953 | json:{"login": "yupengtang", "id": 85978465, "node_id": "MDQ6VXNlcjg1OTc4NDY1", "avatar_url": "https://avatars.githubusercontent.com/u/85978465?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yupengtang", "html_url": "https://github.com/yupengtang", "followers_url": "https://api.github.com/users/yupengtan... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8952/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8952 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8952/timeline | json:[] | value:6191600799 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8952 | value:Shard batched map on batch boundaries with num_proc | value:2026-11-03T21:38:17Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8952/events | value:False | value:Fixes #8919.
With `num_proc > 1`, `Dataset.map` splits the dataset into equal contiguous shards with `shard(contiguous=True)`, ignoring `batch_size`. Every shard then ends with its own incomplete batch:
- a batched function receives short batches in the middle of the dataset, not only at the end;
- `Dataset.bat... | value:https://api.github.com/repos/huggingface/datasets/issues/8952/comments | value:PR_kwDODunzps8AAAABMqjtEQ | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8952", "html_url": "https://github.com/huggingface/datasets/pull/8952", "diff_url": "https://github.com/huggingface/datasets/pull/8952.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8952.patch", "merged_at": null} | value:2026-11-05T17:01:32Z | value:https://api.github.com/repos/huggingface/datasets/issues/8952/labels{/name} | value:8952 | json:{"login": "udsy19", "id": 71968569, "node_id": "MDQ6VXNlcjcxOTY4NTY5", "avatar_url": "https://avatars.githubusercontent.com/u/71968569?v=4", "gravatar_id": "", "url": "https://api.github.com/users/udsy19", "html_url": "https://github.com/udsy19", "followers_url": "https://api.github.com/users/udsy19/followers", "f... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8951/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8951 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8951/timeline | json:[] | value:6191078323 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8951 | value:Fix reading a column that is not in the format columns | value:2026-11-03T21:01:50Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8951/events | value:False | value:After `set_format(type=..., columns=[...])` or `with_transform(..., columns=[...])`, reading a column that is not in the format columns raises `KeyError`:
```python
from datasets import Dataset
ds = Dataset.from_dict({"input_ids": [[1, 2], [3, 4]], "label": [0, 1], "text": ["a b", "c d"]})
ds.set_format(type="n... | value:https://api.github.com/repos/huggingface/datasets/issues/8951/comments | value:PR_kwDODunzps8AAAABMqIx-w | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8951", "html_url": "https://github.com/huggingface/datasets/pull/8951", "diff_url": "https://github.com/huggingface/datasets/pull/8951.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8951.patch", "merged_at": null} | value:2026-11-05T18:32:55Z | value:https://api.github.com/repos/huggingface/datasets/issues/8951/labels{/name} | value:8951 | json:{"login": "akanyaani", "id": 11317416, "node_id": "MDQ6VXNlcjExMzE3NDE2", "avatar_url": "https://avatars.githubusercontent.com/u/11317416?v=4", "gravatar_id": "", "url": "https://api.github.com/users/akanyaani", "html_url": "https://github.com/akanyaani", "followers_url": "https://api.github.com/users/akanyaani/fo... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8950/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8950 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8950/timeline | json:[] | value:6190774085 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8950 | value:Validate metadata.jsonl tables in folder-based builders | value:2026-11-03T20:41:12Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8950/events | value:False | value:What does this PR do?
pyarrow's JSON reader can return a table with invalid list offsets when a list starts with `null`. The `.jsonl` branch of `_read_metadata` in folder-based builders (`imagefolder` and the others) used that table without checking it, so values were silently wrong: `[null, 1]` and `[2]` loaded... | value:https://api.github.com/repos/huggingface/datasets/issues/8950/comments | value:PR_kwDODunzps8AAAABMp5CLw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8950", "html_url": "https://github.com/huggingface/datasets/pull/8950", "diff_url": "https://github.com/huggingface/datasets/pull/8950.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8950.patch", "merged_at": null} | value:2026-11-03T20:41:12Z | value:https://api.github.com/repos/huggingface/datasets/issues/8950/labels{/name} | value:8950 | json:{"login": "shashank11yadav", "id": 65058495, "node_id": "MDQ6VXNlcjY1MDU4NDk1", "avatar_url": "https://avatars.githubusercontent.com/u/65058495?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashank11yadav", "html_url": "https://github.com/shashank11yadav", "followers_url": "https://api.github.com/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8949/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8949 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8949/timeline | json:[] | value:6189886287 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8949 | value:imagefolder metadata.jsonl: a list starting with null loads as wrong values | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-03T19:42:12Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8949/events | value:False | value:### Describe the bug
When the `metadata.jsonl` of an `imagefolder` has a list column whose first element is `null`, `load_dataset` returns wrong values with no error. In the example below, `[null, 1]` and `[2]` come back as `[1, 2]` and `[0]`.
The same rows loaded with the `json` builder come back correctly on ... | value:https://api.github.com/repos/huggingface/datasets/issues/8949/comments | value:I_kwDODunzps8AAAABcPIrTw | value:open | value:1 | value:2026-11-03T19:56:57Z | value:https://api.github.com/repos/huggingface/datasets/issues/8949/labels{/name} | value:8949 | json:{"login": "shashank11yadav", "id": 65058495, "node_id": "MDQ6VXNlcjY1MDU4NDk1", "avatar_url": "https://avatars.githubusercontent.com/u/65058495?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashank11yadav", "html_url": "https://github.com/shashank11yadav", "followers_url": "https://api.github.com/... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8948/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8948 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8948/timeline | json:[] | value:6188863008 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8948 | value:Encode single channel (H, W, 1) arrays with the Image feature | value:2026-11-03T18:35:57Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8948/events | value:False | value:`Image().encode_example` fails on single-channel arrays shaped `(H, W, 1)`, which is how many image pipelines (OpenCV, torchvision, `np.expand_dims`) hand over grayscale images:
```python
import numpy as np
from datasets import Dataset, Features, Image
array = np.zeros((4, 4, 1), dtype=np.uint8)
Dataset.from_di... | value:https://api.github.com/repos/huggingface/datasets/issues/8948/comments | value:PR_kwDODunzps8AAAABMoWKjQ | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8948", "html_url": "https://github.com/huggingface/datasets/pull/8948", "diff_url": "https://github.com/huggingface/datasets/pull/8948.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8948.patch", "merged_at": null} | value:2026-11-05T16:55:39Z | value:https://api.github.com/repos/huggingface/datasets/issues/8948/labels{/name} | value:8948 | json:{"login": "Fahmid-Arman", "id": 184666560, "node_id": "U_kgDOCwHJwA", "avatar_url": "https://avatars.githubusercontent.com/u/184666560?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Fahmid-Arman", "html_url": "https://github.com/Fahmid-Arman", "followers_url": "https://api.github.com/users/Fahmid-Ar... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8947/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8947 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8947/timeline | json:[] | value:6186607240 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8947 | value:Preserve declared scalar types when reading CSV features | value:2026-11-03T16:17:41Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8947/events | value:False | value:CSV readers can reject missing values in declared integer or boolean columns before applying their nullable Arrow schema. With the Python parser, a value such as `"001"` in a declared string column can become `"1"`, losing its leading zeros even though the resulting feature type remains `string`.
Map declared ... | value:https://api.github.com/repos/huggingface/datasets/issues/8947/comments | value:PR_kwDODunzps8AAAABMmhLug | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8947", "html_url": "https://github.com/huggingface/datasets/pull/8947", "diff_url": "https://github.com/huggingface/datasets/pull/8947.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8947.patch", "merged_at": null} | value:2026-11-04T06:38:14Z | value:https://api.github.com/repos/huggingface/datasets/issues/8947/labels{/name} | value:8947 | json:{"login": "Wisnflo", "id": 189682482, "node_id": "U_kgDOC05TMg", "avatar_url": "https://avatars.githubusercontent.com/u/189682482?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Wisnflo", "html_url": "https://github.com/Wisnflo", "followers_url": "https://api.github.com/users/Wisnflo/followers", "fol... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8933/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8933 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8933/timeline | json:[] | value:6171645382 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8933 | value:Fix async callable callbacks in map and filter | value:2026-11-02T21:50:26Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8933/events | value:False | value:`map()` and `filter()` recognize async functions but classify objects with an async `__call__` as synchronous callbacks. A callable mapper then receives a coroutine instead of a mapping; a callable predicate can silently keep every row without being awaited.
Use a shared callback classifier for `Dataset` and `It... | value:https://api.github.com/repos/huggingface/datasets/issues/8933/comments | value:PR_kwDODunzps8AAAABMabj8w | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8933", "html_url": "https://github.com/huggingface/datasets/pull/8933", "diff_url": "https://github.com/huggingface/datasets/pull/8933.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8933.patch", "merged_at": null} | value:2026-11-02T21:50:26Z | value:https://api.github.com/repos/huggingface/datasets/issues/8933/labels{/name} | value:8933 | json:{"login": "GoldArowana", "id": 27399349, "node_id": "MDQ6VXNlcjI3Mzk5MzQ5", "avatar_url": "https://avatars.githubusercontent.com/u/27399349?v=4", "gravatar_id": "", "url": "https://api.github.com/users/GoldArowana", "html_url": "https://github.com/GoldArowana", "followers_url": "https://api.github.com/users/GoldAr... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8932/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8932 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8932/timeline | json:[] | value:6169556813 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8932 | value:Require dill>=0.3.6 on Python 3.11+ | value:2026-11-02T19:33:47Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8932/events | value:False | value:Related to #7537.
`dill` versions before 0.3.6 can't handle code objects on Python 3.11+, so `Dataset.map(num_proc=...)` crashes. `setup.py` still allows `dill>=0.3.0` on every Python version, so an environment that keeps an older `dill` installs `datasets` without complaint and fails at runtime. #7537 shows the... | value:https://api.github.com/repos/huggingface/datasets/issues/8932/comments | value:PR_kwDODunzps8AAAABMYwSjA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8932", "html_url": "https://github.com/huggingface/datasets/pull/8932", "diff_url": "https://github.com/huggingface/datasets/pull/8932.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8932.patch", "merged_at": null} | value:2026-11-02T19:33:47Z | value:https://api.github.com/repos/huggingface/datasets/issues/8932/labels{/name} | value:8932 | json:{"login": "behroozazarkhalili", "id": 80390531, "node_id": "MDQ6VXNlcjgwMzkwNTMx", "avatar_url": "https://avatars.githubusercontent.com/u/80390531?v=4", "gravatar_id": "", "url": "https://api.github.com/users/behroozazarkhalili", "html_url": "https://github.com/behroozazarkhalili", "followers_url": "https://api.gi... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8931/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8931 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8931/timeline | json:[] | value:6163474405 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8931 | value:Fix bare filenames for indices caches | value:2026-11-02T13:26:22Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8931/events | value:False | value:Passing a current-directory filename such as `indices_cache_file_name="selected.arrow"` raises `FileNotFoundError` for a noncontiguous selection: the indices writer calls `os.makedirs("")`. The same path is used by `sort`, `shuffle`, and `train_test_split`.
Use the current directory when the filename has no pa... | value:https://api.github.com/repos/huggingface/datasets/issues/8931/comments | value:PR_kwDODunzps8AAAABMT1xKQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8931", "html_url": "https://github.com/huggingface/datasets/pull/8931", "diff_url": "https://github.com/huggingface/datasets/pull/8931.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8931.patch", "merged_at": null} | value:2026-11-02T13:26:22Z | value:https://api.github.com/repos/huggingface/datasets/issues/8931/labels{/name} | value:8931 | json:{"login": "Linxiushen", "id": 142557582, "node_id": "U_kgDOCH9Bjg", "avatar_url": "https://avatars.githubusercontent.com/u/142557582?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Linxiushen", "html_url": "https://github.com/Linxiushen", "followers_url": "https://api.github.com/users/Linxiushen/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8930/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8930 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8930/timeline | json:[] | value:6162565774 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8930 | value:Preserve `_type` feature columns in dataset card metadata | value:2026-11-02T12:28:51Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8930/events | value:False | value:## Summary
`_type` is a supported feature column name, but serializing it to dataset card metadata raises `TypeError: expected string or bytes-like object, got 'dict'`. The YAML serializer unconditionally removes `_type`, then treats its feature dictionary as a dataclass type name.
```python
from datasets impor... | value:https://api.github.com/repos/huggingface/datasets/issues/8930/comments | value:PR_kwDODunzps8AAAABMTHbSw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8930", "html_url": "https://github.com/huggingface/datasets/pull/8930", "diff_url": "https://github.com/huggingface/datasets/pull/8930.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8930.patch", "merged_at": null} | value:2026-11-02T12:39:25Z | value:https://api.github.com/repos/huggingface/datasets/issues/8930/labels{/name} | value:8930 | json:{"login": "linjiaye929", "id": 91742523, "node_id": "U_kgDOBXfhOw", "avatar_url": "https://avatars.githubusercontent.com/u/91742523?v=4", "gravatar_id": "", "url": "https://api.github.com/users/linjiaye929", "html_url": "https://github.com/linjiaye929", "followers_url": "https://api.github.com/users/linjiaye929/fo... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8929/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8929 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8929/timeline | json:[] | value:6155715894 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8929 | value:Fix non-batched Dataset.filter with arrow and polars formats | value:2026-11-02T03:37:21Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8929/events | value:False | value:With the `arrow` or `polars` format, a non-batched `Dataset.filter` crashes before it calls the function even once:
```python
import pyarrow.compute as pc
from datasets import Dataset
ds = Dataset.from_dict({"a": list(range(10))})
ds.with_format("polars").filter(lambda df: df["a"] > 5)
# AttributeError: 'DataF... | value:https://api.github.com/repos/huggingface/datasets/issues/8929/comments | value:PR_kwDODunzps8AAAABMNoKtg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8929", "html_url": "https://github.com/huggingface/datasets/pull/8929", "diff_url": "https://github.com/huggingface/datasets/pull/8929.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8929.patch", "merged_at": null} | value:2026-11-02T03:37:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8929/labels{/name} | value:8929 | json:{"login": "astroknot-sheep", "id": 105595392, "node_id": "U_kgDOBktCAA", "avatar_url": "https://avatars.githubusercontent.com/u/105595392?v=4", "gravatar_id": "", "url": "https://api.github.com/users/astroknot-sheep", "html_url": "https://github.com/astroknot-sheep", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8928/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8928 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8928/timeline | json:[] | value:6155664033 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8928 | value:Keep string values of mixed-type JSON fields instead of parsing them as JSON | value:2026-11-02T03:31:37Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8928/events | value:False | value:Fixes #8926.
When a field has mixed types, the JSON loader stores it as `Json()`. Before encoding a value, `json_encode_field` tries `ujson_loads` on it and keeps it unchanged if that succeeds, assuming it is already JSON-encoded. So a string that happens to be valid JSON text is stored without its quotes and co... | value:https://api.github.com/repos/huggingface/datasets/issues/8928/comments | value:PR_kwDODunzps8AAAABMNli8A | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8928", "html_url": "https://github.com/huggingface/datasets/pull/8928", "diff_url": "https://github.com/huggingface/datasets/pull/8928.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8928.patch", "merged_at": null} | value:2026-11-02T03:31:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8928/labels{/name} | value:8928 | json:{"login": "behroozazarkhalili", "id": 80390531, "node_id": "MDQ6VXNlcjgwMzkwNTMx", "avatar_url": "https://avatars.githubusercontent.com/u/80390531?v=4", "gravatar_id": "", "url": "https://api.github.com/users/behroozazarkhalili", "html_url": "https://github.com/behroozazarkhalili", "followers_url": "https://api.gi... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8927/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8927 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8927/timeline | json:[] | value:6154387424 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8927 | value:Preserve nested fields when combining streaming datasets | value:2026-11-02T01:15:59Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8927/events | value:False | value:## Preserve nested fields when combining streaming datasets
Combining generator-backed streams with different finite struct fields silently drops fields present only in later sources. Arrow-backed streams fail to cast the same inputs, while map-style concatenation already preserves their union.
Recursively comb... | value:https://api.github.com/repos/huggingface/datasets/issues/8927/comments | value:PR_kwDODunzps8AAAABMMlIpw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8927", "html_url": "https://github.com/huggingface/datasets/pull/8927", "diff_url": "https://github.com/huggingface/datasets/pull/8927.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8927.patch", "merged_at": null} | value:2026-11-02T01:15:59Z | value:https://api.github.com/repos/huggingface/datasets/issues/8927/labels{/name} | value:8927 | json:{"login": "Aurantium-47", "id": 280805072, "node_id": "U_kgDOELy-0A", "avatar_url": "https://avatars.githubusercontent.com/u/280805072?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Aurantium-47", "html_url": "https://github.com/Aurantium-47", "followers_url": "https://api.github.com/users/Aurantium... | value:True | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8926/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8926 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8926/timeline | json:[] | value:6154203695 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8926 | value:JSON loader turns string values into numbers, booleans or null when a field has mixed types | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-11-02T00:56:45Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8926/events | value:False | value:### Describe the bug
When a field has mixed types, the JSON loader stores it as `Json()` (the default `on_mixed_types="use_json"` from #8027). Any string value in that field that is itself valid JSON text is then stored without its quotes, so it comes back as a different value and type: `"123"` becomes `123`, `"... | value:https://api.github.com/repos/huggingface/datasets/issues/8926/comments | value:I_kwDODunzps8AAAABbtGyLw | value:open | value:0 | value:2026-11-02T00:56:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8926/labels{/name} | value:8926 | json:{"login": "behroozazarkhalili", "id": 80390531, "node_id": "MDQ6VXNlcjgwMzkwNTMx", "avatar_url": "https://avatars.githubusercontent.com/u/80390531?v=4", "gravatar_id": "", "url": "https://api.github.com/users/behroozazarkhalili", "html_url": "https://github.com/behroozazarkhalili", "followers_url": "https://api.gi... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8925/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8925 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8925/timeline | json:[] | value:6152845737 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8925 | value:Raise in cast_column when the column is not in the dataset | value:2026-11-01T22:27:40Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8925/events | value:False | value:`cast_column` with a column name that is not in the dataset adds a new column filled with `None` instead of raising, in two places:
```python
from datasets import Dataset, Image, Value
ds = Dataset.from_dict({"a": [1, 2]})
ds.cast_column("b", Image())[0]
# {'a': 1, 'b': None}
ids = ds.to_iterable_dataset()
lis... | value:https://api.github.com/repos/huggingface/datasets/issues/8925/comments | value:PR_kwDODunzps8AAAABMLX-wA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8925", "html_url": "https://github.com/huggingface/datasets/pull/8925", "diff_url": "https://github.com/huggingface/datasets/pull/8925.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8925.patch", "merged_at": null} | value:2026-11-01T22:27:40Z | value:https://api.github.com/repos/huggingface/datasets/issues/8925/labels{/name} | value:8925 | json:{"login": "ANDROIDHASSAN", "id": 84630851, "node_id": "MDQ6VXNlcjg0NjMwODUx", "avatar_url": "https://avatars.githubusercontent.com/u/84630851?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ANDROIDHASSAN", "html_url": "https://github.com/ANDROIDHASSAN", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8924/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8924 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8924/timeline | json:[] | value:6151767736 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8924 | value:Close decoder-owned streams when decoding remote PDFs | value:2026-11-01T20:35:52Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8924/events | value:False | value:## Summary
Source inspection found that `Pdf.decode_example()` opens remote streams itself but passes them to `pdfplumber.open()` as externally owned file-like objects. Consequently, closing the returned PDF (including leaving its context manager) does not close the decoder's stream, and constructor failures als... | value:https://api.github.com/repos/huggingface/datasets/issues/8924/comments | value:PR_kwDODunzps8AAAABMKiRuw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8924", "html_url": "https://github.com/huggingface/datasets/pull/8924", "diff_url": "https://github.com/huggingface/datasets/pull/8924.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8924.patch", "merged_at": null} | value:2026-11-01T21:27:26Z | value:https://api.github.com/repos/huggingface/datasets/issues/8924/labels{/name} | value:8924 | json:{"login": "Dodothereal", "id": 129273127, "node_id": "U_kgDOB7SNJw", "avatar_url": "https://avatars.githubusercontent.com/u/129273127?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Dodothereal", "html_url": "https://github.com/Dodothereal", "followers_url": "https://api.github.com/users/Dodothereal/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8923/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8923 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8923/timeline | json:[] | value:6151521578 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8923 | value:Fix double nesting of list subfeatures in Sequence(dict) | value:2026-11-01T20:12:52Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8923/events | value:False | value:### Fix double nesting of list subfeatures in `Sequence(dict)`
In the TFDS-compat conversion, `Sequence(feature={...})` turns a dict of subfeatures into a dict of lists. Since #7672, subfeatures that are already list features are not converted anymore **when all of them are lists** — but the mixed case was still... | value:https://api.github.com/repos/huggingface/datasets/issues/8923/comments | value:PR_kwDODunzps8AAAABMKVu9g | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8923", "html_url": "https://github.com/huggingface/datasets/pull/8923", "diff_url": "https://github.com/huggingface/datasets/pull/8923.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8923.patch", "merged_at": null} | value:2026-11-01T20:13:01Z | value:https://api.github.com/repos/huggingface/datasets/issues/8923/labels{/name} | value:8923 | json:{"login": "simpleqt", "id": 89645338, "node_id": "MDQ6VXNlcjg5NjQ1MzM4", "avatar_url": "https://avatars.githubusercontent.com/u/89645338?v=4", "gravatar_id": "", "url": "https://api.github.com/users/simpleqt", "html_url": "https://github.com/simpleqt", "followers_url": "https://api.github.com/users/simpleqt/follow... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8922/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8922 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8922/timeline | json:[] | value:6145068682 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8922 | value:Fix ClassLabel.int2str for one-shot iterables | value:2026-11-01T09:50:08Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8922/events | value:False | value:## Summary
`ClassLabel.int2str` accepts `Iterable` inputs, but currently iterates over them once to validate the labels and again to convert them. A one-shot iterator or generator is exhausted by validation, so valid input silently returns an empty list:
```python
label = ClassLabel(names=["negative", "pos... | value:https://api.github.com/repos/huggingface/datasets/issues/8922/comments | value:PR_kwDODunzps8AAAABMFVKyQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8922", "html_url": "https://github.com/huggingface/datasets/pull/8922", "diff_url": "https://github.com/huggingface/datasets/pull/8922.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8922.patch", "merged_at": null} | value:2026-11-01T09:54:16Z | value:https://api.github.com/repos/huggingface/datasets/issues/8922/labels{/name} | value:8922 | json:{"login": "linjiaye929", "id": 91742523, "node_id": "U_kgDOBXfhOw", "avatar_url": "https://avatars.githubusercontent.com/u/91742523?v=4", "gravatar_id": "", "url": "https://api.github.com/users/linjiaye929", "html_url": "https://github.com/linjiaye929", "followers_url": "https://api.github.com/users/linjiaye929/fo... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8921/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8921 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8921/timeline | json:[] | value:6140673659 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8921 | value:Fix dead links to WebDataset, Spark and Albumentations docs | value:2026-10-31T22:50:50Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8921/events | value:False | value:Five links in the docs returned 404. Updated to pages that exist (each new URL returns 200):
- `webdataset.github.io/webdataset` (audio, image, video dataset guides) → `github.com/webdataset/webdataset`: the docs site is gone and the project moved from `tmbdev` to the `webdataset` org.
- `use_with_spark`: Spark ... | value:https://api.github.com/repos/huggingface/datasets/issues/8921/comments | value:PR_kwDODunzps8AAAABMB6eFA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8921", "html_url": "https://github.com/huggingface/datasets/pull/8921", "diff_url": "https://github.com/huggingface/datasets/pull/8921.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8921.patch", "merged_at": null} | value:2026-10-31T22:50:50Z | value:https://api.github.com/repos/huggingface/datasets/issues/8921/labels{/name} | value:8921 | json:{"login": "erdogan-deniz", "id": 68436196, "node_id": "MDQ6VXNlcjY4NDM2MTk2", "avatar_url": "https://avatars.githubusercontent.com/u/68436196?v=4", "gravatar_id": "", "url": "https://api.github.com/users/erdogan-deniz", "html_url": "https://github.com/erdogan-deniz", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8920/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8920 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8920/timeline | json:[] | value:6140128360 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8920 | value:Preserve concatenated table rows when pickling integer column selections | value:2026-10-31T21:48:10Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8920/events | value:False | value:`ConcatenationTable.select([1])` returns the expected table immediately, but discards the selected columns from its stored blocks. Serializing the result then silently loses its rows, including after slicing or filtering.
Resolve the validated selection to column names before selecting the stored blocks. This pr... | value:https://api.github.com/repos/huggingface/datasets/issues/8920/comments | value:PR_kwDODunzps8AAAABMBe84w | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8920", "html_url": "https://github.com/huggingface/datasets/pull/8920", "diff_url": "https://github.com/huggingface/datasets/pull/8920.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8920.patch", "merged_at": null} | value:2026-10-31T21:48:10Z | value:https://api.github.com/repos/huggingface/datasets/issues/8920/labels{/name} | value:8920 | json:{"login": "ElMagAmine", "id": 36851541, "node_id": "MDQ6VXNlcjM2ODUxNTQx", "avatar_url": "https://avatars.githubusercontent.com/u/36851541?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ElMagAmine", "html_url": "https://github.com/ElMagAmine", "followers_url": "https://api.github.com/users/ElMagAmin... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8919/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "IlyaKhalafi", "id": 80855937, "node_id": "MDQ6VXNlcjgwODU1OTM3", "avatar_url": "https://avatars.githubusercontent.com/u/80855937?v=4", "gravatar_id": "", "url": "https://api.github.com/users/IlyaKhalafi", "html_url": "https://github.com/IlyaKhalafi", "followers_url": "https://api.github.com/users/IlyaKh... | value:https://github.com/huggingface/datasets/issues/8919 | json:[{"login": "IlyaKhalafi", "id": 80855937, "node_id": "MDQ6VXNlcjgwODU1OTM3", "avatar_url": "https://avatars.githubusercontent.com/u/80855937?v=4", "gravatar_id": "", "url": "https://api.github.com/users/IlyaKhalafi", "html_url": "https://github.com/IlyaKhalafi", "followers_url": "https://api.github.com/users/IlyaK... | value:https://api.github.com/repos/huggingface/datasets/issues/8919/timeline | json:[] | value:6139692528 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8919 | value:Batched map / Dataset.batch with num_proc>1 splits batches at shard boundaries and drop_last_batch loses rows | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-31T20:58:24Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8919/events | value:False | value:### Describe the bug
With `num_proc > 1`, batched `Dataset.map` (and `Dataset.batch`, which goes through it) splits the dataset into `num_proc` equal row chunks via `self.shard(num_shards, index=rank, contiguous=True)`, ignoring `batch_size`. Each chunk ends with its own short batch, so the function sees wrongly... | value:https://api.github.com/repos/huggingface/datasets/issues/8919/comments | value:I_kwDODunzps8AAAABbfRF8A | value:open | value:3 | value:2026-11-03T19:45:55Z | value:https://api.github.com/repos/huggingface/datasets/issues/8919/labels{/name} | value:8919 | json:{"login": "udsy19", "id": 71968569, "node_id": "MDQ6VXNlcjcxOTY4NTY5", "avatar_url": "https://avatars.githubusercontent.com/u/71968569?v=4", "gravatar_id": "", "url": "https://api.github.com/users/udsy19", "html_url": "https://github.com/udsy19", "followers_url": "https://api.github.com/users/udsy19/followers", "f... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8918/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8918 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8918/timeline | json:[] | value:6139037909 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8918 | value:Fix map cache reuse across num_proc when the cache path has glob/regex special characters | value:2026-10-31T19:44:36Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8918/events | value:False | value:`Dataset.map` reuses cache files written with a different `num_proc` (#7434) by globbing `f"{cache_file_prefix}*{cache_file_ext}"` and parsing each match with `string_to_dict(path, prefix + suffix_template + ext)`. The prefix is a full filesystem path used unescaped as both a glob and a regex, so special characte... | value:https://api.github.com/repos/huggingface/datasets/issues/8918/comments | value:PR_kwDODunzps8AAAABMAo4pQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8918", "html_url": "https://github.com/huggingface/datasets/pull/8918", "diff_url": "https://github.com/huggingface/datasets/pull/8918.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8918.patch", "merged_at": null} | value:2026-10-31T19:44:36Z | value:https://api.github.com/repos/huggingface/datasets/issues/8918/labels{/name} | value:8918 | json:{"login": "udsy19", "id": 71968569, "node_id": "MDQ6VXNlcjcxOTY4NTY5", "avatar_url": "https://avatars.githubusercontent.com/u/71968569?v=4", "gravatar_id": "", "url": "https://api.github.com/users/udsy19", "html_url": "https://github.com/udsy19", "followers_url": "https://api.github.com/users/udsy19/followers", "f... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8917/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8917 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8917/timeline | json:[] | value:6138782804 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8917 | value:Fix casting a nullable LargeList of ArrayXD to a fixed-length List | value:2026-10-31T19:18:31Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8917/events | value:False | value:Casting a nullable `LargeList` of an ArrayXD feature to a fixed-length `List` crashes, while the same cast from a regular `List` works:
```python
from datasets import Dataset, Features, Array2D, List, LargeList
ds = Dataset.from_dict(
{"x": [[[[1, 2]], [[3, 4]]], None]},
features=Features({"x": LargeLis... | value:https://api.github.com/repos/huggingface/datasets/issues/8917/comments | value:PR_kwDODunzps8AAAABMAcO9g | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8917", "html_url": "https://github.com/huggingface/datasets/pull/8917", "diff_url": "https://github.com/huggingface/datasets/pull/8917.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8917.patch", "merged_at": null} | value:2026-10-31T19:18:31Z | value:https://api.github.com/repos/huggingface/datasets/issues/8917/labels{/name} | value:8917 | json:{"login": "udsy19", "id": 71968569, "node_id": "MDQ6VXNlcjcxOTY4NTY5", "avatar_url": "https://avatars.githubusercontent.com/u/71968569?v=4", "gravatar_id": "", "url": "https://api.github.com/users/udsy19", "html_url": "https://github.com/udsy19", "followers_url": "https://api.github.com/users/udsy19/followers", "f... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8916/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8916 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8916/timeline | json:[] | value:6138436855 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8916 | value:Fix default configuration selection when updating dataset cards | value:2026-10-31T18:41:21Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8916/events | value:False | value:Setting `set_default=True` during a push can either raise `KeyError('default')` or restore stale `data_files` in the dataset card.
The card writer treats a sole configuration as the default even when it has no explicit `default` flag, then unconditionally pops that missing flag. If the selected configuration is ... | value:https://api.github.com/repos/huggingface/datasets/issues/8916/comments | value:PR_kwDODunzps8AAAABMALEmQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8916", "html_url": "https://github.com/huggingface/datasets/pull/8916", "diff_url": "https://github.com/huggingface/datasets/pull/8916.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8916.patch", "merged_at": null} | value:2026-10-31T18:41:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8916/labels{/name} | value:8916 | json:{"login": "DiadkoShmek", "id": 207617272, "node_id": "U_kgDODF_8-A", "avatar_url": "https://avatars.githubusercontent.com/u/207617272?v=4", "gravatar_id": "", "url": "https://api.github.com/users/DiadkoShmek", "html_url": "https://github.com/DiadkoShmek", "followers_url": "https://api.github.com/users/DiadkoShmek/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8915/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8915 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8915/timeline | json:[] | value:6133968824 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8915 | value:fix: encode non-JSON strings per value in Json.cast_storage | value:2026-10-31T10:49:24Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8915/events | value:False | value:Fixes #8914
## What this fixes
`Json.cast_storage` parsed only `storage[:5]` to decide whether a string column already contains JSON. When those five values parsed but a later one did not, the whole column was left untouched: the later value was stored as raw text instead of being encoded as a JSON string, and ... | value:https://api.github.com/repos/huggingface/datasets/issues/8915/comments | value:PR_kwDODunzps8AAAABL8uavA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8915", "html_url": "https://github.com/huggingface/datasets/pull/8915", "diff_url": "https://github.com/huggingface/datasets/pull/8915.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8915.patch", "merged_at": null} | value:2026-10-31T10:49:24Z | value:https://api.github.com/repos/huggingface/datasets/issues/8915/labels{/name} | value:8915 | json:{"login": "Steve-20-ai", "id": 306727257, "node_id": "U_kgDOEkhJWQ", "avatar_url": "https://avatars.githubusercontent.com/u/306727257?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Steve-20-ai", "html_url": "https://github.com/Steve-20-ai", "followers_url": "https://api.github.com/users/Steve-20-ai/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8914/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "Steve-20-ai", "id": 306727257, "node_id": "U_kgDOEkhJWQ", "avatar_url": "https://avatars.githubusercontent.com/u/306727257?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Steve-20-ai", "html_url": "https://github.com/Steve-20-ai", "followers_url": "https://api.github.com/users/Steve-20-ai/... | value:https://github.com/huggingface/datasets/issues/8914 | json:[{"login": "Steve-20-ai", "id": 306727257, "node_id": "U_kgDOEkhJWQ", "avatar_url": "https://avatars.githubusercontent.com/u/306727257?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Steve-20-ai", "html_url": "https://github.com/Steve-20-ai", "followers_url": "https://api.github.com/users/Steve-20-ai... | value:https://api.github.com/repos/huggingface/datasets/issues/8914/timeline | json:[] | value:6133955166 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8914 | value:Json.cast_storage only inspects the first 5 values, so plain strings past that point are stored unencoded and fail to decode | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-31T10:47:45Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8914/events | value:False | value:### Describe the bug
Casting a string column to `Json()` uses a 5-row sample to decide whether the column already holds JSON. If the first five values parse as JSON but a later value does not, the later value is stored **unencoded**. It then reads back through `decode_example` as a syntax error instead of as a s... | value:https://api.github.com/repos/huggingface/datasets/issues/8914/comments | value:I_kwDODunzps8AAAABbZy6Xg | value:open | value:1 | value:2026-10-31T10:48:03Z | value:https://api.github.com/repos/huggingface/datasets/issues/8914/labels{/name} | value:8914 | json:{"login": "Steve-20-ai", "id": 306727257, "node_id": "U_kgDOEkhJWQ", "avatar_url": "https://avatars.githubusercontent.com/u/306727257?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Steve-20-ai", "html_url": "https://github.com/Steve-20-ai", "followers_url": "https://api.github.com/users/Steve-20-ai/... | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8913/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8913 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8913/timeline | json:[] | value:6129008846 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8913 | value:Fix empty typed Dataset save and reload | value:2026-10-31T00:15:19Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8913/events | value:False | value:Saving an empty typed Dataset with default sharding caps the automatically chosen shard count at zero. The save completes, but `load_from_disk()` cannot reconstruct a table from the empty file list.
Keep at least one automatically chosen shard. Add regression coverage for default settings, `num_proc=2` and `max_... | value:https://api.github.com/repos/huggingface/datasets/issues/8913/comments | value:PR_kwDODunzps8AAAABL40z5A | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8913", "html_url": "https://github.com/huggingface/datasets/pull/8913", "diff_url": "https://github.com/huggingface/datasets/pull/8913.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8913.patch", "merged_at": null} | value:2026-11-09T17:53:27Z | value:https://api.github.com/repos/huggingface/datasets/issues/8913/labels{/name} | value:8913 | json:{"login": "alberttwong", "id": 749093, "node_id": "MDQ6VXNlcjc0OTA5Mw==", "avatar_url": "https://avatars.githubusercontent.com/u/749093?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alberttwong", "html_url": "https://github.com/alberttwong", "followers_url": "https://api.github.com/users/alberttwon... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8912/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8912 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8912/timeline | json:[] | value:6128941306 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8912 | value:Empty typed Dataset default save cannot reload because it writes zero shards | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-31T00:07:45Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8912/events | value:False | value:### Describe the bug
Datasets 5.1.0 writes zero shards when saving an empty typed Dataset with default sharding. The saved state has `_data_files: []`; `load_from_disk()` raises `IndexError: list index out of range` in `ConcatenationTable.from_tables`. This reproduces without Redis, ADBC, network access or downl... | value:https://api.github.com/repos/huggingface/datasets/issues/8912/comments | value:I_kwDODunzps8AAAABbVA4-g | value:open | value:0 | value:2026-10-31T00:07:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8912/labels{/name} | value:8912 | json:{"login": "alberttwong", "id": 749093, "node_id": "MDQ6VXNlcjc0OTA5Mw==", "avatar_url": "https://avatars.githubusercontent.com/u/749093?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alberttwong", "html_url": "https://github.com/alberttwong", "followers_url": "https://api.github.com/users/alberttwon... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8911/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8911 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8911/timeline | json:[] | value:6128926592 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8911 | value:docs: link Redis ADBC Dataset integration guide | value:2026-10-31T00:06:22Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8911/events | value:False | value:Link a concrete Redis SQL-to-Arrow-to-Dataset workflow from the PyArrow interoperability guide. Explain the external driver/Search prerequisite, local conversion/persistence scope and one-shard empty-result save setting. No built-in loader or dependency is added.
Validation: the exact linked guide example passed... | value:https://api.github.com/repos/huggingface/datasets/issues/8911/comments | value:PR_kwDODunzps8AAAABL4wpAA | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8911", "html_url": "https://github.com/huggingface/datasets/pull/8911", "diff_url": "https://github.com/huggingface/datasets/pull/8911.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8911.patch", "merged_at": null} | value:2026-11-09T17:51:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8911/labels{/name} | value:8911 | json:{"login": "alberttwong", "id": 749093, "node_id": "MDQ6VXNlcjc0OTA5Mw==", "avatar_url": "https://avatars.githubusercontent.com/u/749093?v=4", "gravatar_id": "", "url": "https://api.github.com/users/alberttwong", "html_url": "https://github.com/alberttwong", "followers_url": "https://api.github.com/users/alberttwon... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8910/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8910 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8910/timeline | json:[] | value:6127372052 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8910 | value:Docs: add missing imports in Use with Polars data extraction example | value:2026-10-30T21:40:09Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8910/events | value:False | value:Fixes #8909
The "Example: data extraction" code block in use_with_polars.mdx used `re` and `pl` without importing them, so it failed with a NameError when run on its own. This adds `import re` and `import polars as pl` at the top of the block. | value:https://api.github.com/repos/huggingface/datasets/issues/8910/comments | value:PR_kwDODunzps8AAAABL3g9yw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8910", "html_url": "https://github.com/huggingface/datasets/pull/8910", "diff_url": "https://github.com/huggingface/datasets/pull/8910.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8910.patch", "merged_at": null} | value:2026-10-30T21:40:09Z | value:https://api.github.com/repos/huggingface/datasets/issues/8910/labels{/name} | value:8910 | json:{"login": "punit857", "id": 192017818, "node_id": "U_kgDOC3H1mg", "avatar_url": "https://avatars.githubusercontent.com/u/192017818?v=4", "gravatar_id": "", "url": "https://api.github.com/users/punit857", "html_url": "https://github.com/punit857", "followers_url": "https://api.github.com/users/punit857/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8909/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8909 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8909/timeline | json:[] | value:6127028924 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8909 | value:Docs: missing `import re` in "Use with Polars" data extraction example | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-30T21:12:39Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8909/events | value:False | value:### Describe the bug
The "Example: data extraction" code block on https://huggingface.co/docs/datasets/en/use_with_polars uses `re` without importing it:
pattern = re.compile("boxed\\{(.*)\\}")
Running the block as shown fails with `NameError: name 're' is not defined`.
### Steps to reproduce the bug
Cop... | value:https://api.github.com/repos/huggingface/datasets/issues/8909/comments | value:I_kwDODunzps8AAAABbTMKvA | value:open | value:1 | value:2026-10-30T21:29:27Z | value:https://api.github.com/repos/huggingface/datasets/issues/8909/labels{/name} | value:8909 | json:{"login": "docsalive-team", "id": 350483914, "node_id": "U_kgDOFOP1yg", "avatar_url": "https://avatars.githubusercontent.com/u/350483914?v=4", "gravatar_id": "", "url": "https://api.github.com/users/docsalive-team", "html_url": "https://github.com/docsalive-team", "followers_url": "https://api.github.com/users/doc... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8908/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8908 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8908/timeline | json:[] | value:6109381173 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8908 | value:Close xopen handles in xnumpy_load, xpandas_read_excel, xpyarrow_parquet_read_table, xsio_loadmat | value:2026-10-29T22:11:42Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8908/events | value:False | value:Fixes #8907
Follow-up to #8786, which fixed this for `xpandas_read_csv`. The remaining streaming readers still handed `xopen(...)` to the reader without closing it:
- `xpyarrow_parquet_read_table`, `xsio_loadmat`: wrap in `with xopen(...)`.
- `xpandas_read_excel`: both the direct read and the BytesIO fallback n... | value:https://api.github.com/repos/huggingface/datasets/issues/8908/comments | value:PR_kwDODunzps8AAAABLpILBQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8908", "html_url": "https://github.com/huggingface/datasets/pull/8908", "diff_url": "https://github.com/huggingface/datasets/pull/8908.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8908.patch", "merged_at": null} | value:2026-10-29T22:11:42Z | value:https://api.github.com/repos/huggingface/datasets/issues/8908/labels{/name} | value:8908 | json:{"login": "michaelkillgta", "id": 269775689, "node_id": "U_kgDOEBRzSQ", "avatar_url": "https://avatars.githubusercontent.com/u/269775689?v=4", "gravatar_id": "", "url": "https://api.github.com/users/michaelkillgta", "html_url": "https://github.com/michaelkillgta", "followers_url": "https://api.github.com/users/mic... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8907/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8907 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8907/timeline | json:[] | value:6109377081 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8907 | value:xnumpy_load, xpandas_read_excel, xpyarrow_parquet_read_table and xsio_loadmat never close their xopen handle | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-29T22:11:22Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8907/events | value:False | value:### Describe the bug
#8786 fixed the leaked `xopen` handle in `xpandas_read_csv`, but the other streaming readers in `utils/file_utils.py` still pass `xopen(...)` straight to the reader and never close it:
- `xnumpy_load`
- `xpandas_read_excel` (both the first try and the BytesIO fallback)
- `xpyarrow_parquet_r... | value:https://api.github.com/repos/huggingface/datasets/issues/8907/comments | value:I_kwDODunzps8AAAABbCWyOQ | value:open | value:0 | value:2026-10-29T22:11:22Z | value:https://api.github.com/repos/huggingface/datasets/issues/8907/labels{/name} | value:8907 | json:{"login": "michaelkillgta", "id": 269775689, "node_id": "U_kgDOEBRzSQ", "avatar_url": "https://avatars.githubusercontent.com/u/269775689?v=4", "gravatar_id": "", "url": "https://api.github.com/users/michaelkillgta", "html_url": "https://github.com/michaelkillgta", "followers_url": "https://api.github.com/users/mic... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8905/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8905 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8905/timeline | json:[] | value:6100014532 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8905 | value:Bump doc-builder to get the docs machine-translated | value:2026-10-29T12:50:36Z | value:CONTRIBUTOR | json:{"login": "mishig25", "id": 11827707, "node_id": "MDQ6VXNlcjExODI3NzA3", "avatar_url": "https://avatars.githubusercontent.com/u/11827707?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mishig25", "html_url": "https://github.com/mishig25", "followers_url": "https://api.github.com/users/mishig25/follow... | value:https://api.github.com/repos/huggingface/datasets/issues/8905/events | value:False | value:2026-10-30T14:50:37Z | value:Bumps doc-builder to the latest `main` so the docs on hf.co/docs get [machine-translated](https://github.com/huggingface/doc-builder#machine-translated-documentation) into 11 languages, as discussed on Slack. Pages built by older doc-builder versions can't be translated.
🤖 Generated with [Claude Code](https://c... | value:https://api.github.com/repos/huggingface/datasets/issues/8905/comments | value:PR_kwDODunzps8AAAABLhhhYw | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8905", "html_url": "https://github.com/huggingface/datasets/pull/8905", "diff_url": "https://github.com/huggingface/datasets/pull/8905.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8905.patch", "merged_at": "2026-10-30T14:50:37Z"} | value:2026-10-30T14:50:40Z | value:https://api.github.com/repos/huggingface/datasets/issues/8905/labels{/name} | value:8905 | json:{"login": "mishig25", "id": 11827707, "node_id": "MDQ6VXNlcjExODI3NzA3", "avatar_url": "https://avatars.githubusercontent.com/u/11827707?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mishig25", "html_url": "https://github.com/mishig25", "followers_url": "https://api.github.com/users/mishig25/follow... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8904/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8904 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8904/timeline | json:[] | value:6093849240 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8904 | value:Preserve on-the-fly transforms when flattening dataset indices | value:2026-10-29T05:55:15Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8904/events | value:False | value:Dataset.flatten_indices() currently passes the dataset’s active format to an identity map. With a custom transform, the transform’s output is written to Arrow, while the result retains the same transform and applies it again on access.
For example, for stored values [2, 0] and an on-the-fly transform that adds 1... | value:https://api.github.com/repos/huggingface/datasets/issues/8904/comments | value:PR_kwDODunzps8AAAABLcheZA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8904", "html_url": "https://github.com/huggingface/datasets/pull/8904", "diff_url": "https://github.com/huggingface/datasets/pull/8904.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8904.patch", "merged_at": null} | value:2026-10-30T09:21:02Z | value:https://api.github.com/repos/huggingface/datasets/issues/8904/labels{/name} | value:8904 | json:{"login": "Mixriaoliao", "id": 261539271, "node_id": "U_kgDOD5bFxw", "avatar_url": "https://avatars.githubusercontent.com/u/261539271?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Mixriaoliao", "html_url": "https://github.com/Mixriaoliao", "followers_url": "https://api.github.com/users/Mixriaoliao/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8903/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8903 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8903/timeline | json:[] | value:6080889794 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8903 | value:Load Harbor multi-step tasks without a top-level instruction.md | value:2026-10-28T14:23:32Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8903/events | value:False | value:Title: Load Harbor multi-step tasks without a top-level instruction.md
A multi-step Harbor task declares `[[steps]]` in its `task.toml` and keeps its instructions in `steps/<step>/instruction.md`, so it doesn't need an `instruction.md` next to its `task.toml`. Harbor handles that case in several places:
- `Task... | value:https://api.github.com/repos/huggingface/datasets/issues/8903/comments | value:PR_kwDODunzps8AAAABLSFV4A | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8903", "html_url": "https://github.com/huggingface/datasets/pull/8903", "diff_url": "https://github.com/huggingface/datasets/pull/8903.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8903.patch", "merged_at": null} | value:2026-10-28T14:23:32Z | value:https://api.github.com/repos/huggingface/datasets/issues/8903/labels{/name} | value:8903 | json:{"login": "xlb3055", "id": 178128333, "node_id": "U_kgDOCp4FzQ", "avatar_url": "https://avatars.githubusercontent.com/u/178128333?v=4", "gravatar_id": "", "url": "https://api.github.com/users/xlb3055", "html_url": "https://github.com/xlb3055", "followers_url": "https://api.github.com/users/xlb3055/followers", "fol... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8902/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8902 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8902/timeline | json:[] | value:6078754655 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8902 | value:Allow reading prepared datasets from read-only shared caches | value:2026-10-28T12:17:51Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8902/events | value:False | value:# Allow reading prepared datasets from read-only shared caches
Addresses the Datasets lock-file failure reported in huggingface/datasets#7900. Hub cache permissions are outside this patch.
## Problem
`DatasetBuilder` acquires exclusive file locks while checking cached metadata and while deciding whether a data... | value:https://api.github.com/repos/huggingface/datasets/issues/8902/comments | value:PR_kwDODunzps8AAAABLQXf4A | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8902", "html_url": "https://github.com/huggingface/datasets/pull/8902", "diff_url": "https://github.com/huggingface/datasets/pull/8902.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8902.patch", "merged_at": null} | value:2026-10-28T12:17:51Z | value:https://api.github.com/repos/huggingface/datasets/issues/8902/labels{/name} | value:8902 | json:{"login": "Oscar-man-shrestha", "id": 172575019, "node_id": "U_kgDOCklJKw", "avatar_url": "https://avatars.githubusercontent.com/u/172575019?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Oscar-man-shrestha", "html_url": "https://github.com/Oscar-man-shrestha", "followers_url": "https://api.github.c... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8901/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8901 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8901/timeline | json:[] | value:6075858208 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8901 | value:Use the same class names in every split in DatasetDict.class_encode_column | value:2026-10-28T09:06:02Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8901/events | value:False | value:`DatasetDict.class_encode_column` encoded each split on its own, so a split that lacks some classes got a shorter `ClassLabel`, and the same label got a different integer id than in the other splits. Using the train features to decode the ids of another split then returned the wrong class, and concatenating the s... | value:https://api.github.com/repos/huggingface/datasets/issues/8901/comments | value:PR_kwDODunzps8AAAABLODlpQ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8901", "html_url": "https://github.com/huggingface/datasets/pull/8901", "diff_url": "https://github.com/huggingface/datasets/pull/8901.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8901.patch", "merged_at": null} | value:2026-10-28T09:06:02Z | value:https://api.github.com/repos/huggingface/datasets/issues/8901/labels{/name} | value:8901 | json:{"login": "mottopanikeiku", "id": 176798723, "node_id": "U_kgDOCom8Aw", "avatar_url": "https://avatars.githubusercontent.com/u/176798723?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mottopanikeiku", "html_url": "https://github.com/mottopanikeiku", "followers_url": "https://api.github.com/users/mot... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8900/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8900 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8900/timeline | json:[] | value:6068218472 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8900 | value:Restore parent tokenizer environment after multiprocessing map | value:2026-10-27T21:41:44Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8900/events | value:False | value:`Dataset.map(num_proc=2)` currently leaks its temporary `TOKENIZERS_PARALLELISM=false` override into the parent process. On a cold map, Python's copied `os.environ` reports the old value while a newly launched subprocess still inherits `false`. Fully cached maps and pool-constructor errors leave the Python mappin... | value:https://api.github.com/repos/huggingface/datasets/issues/8900/comments | value:PR_kwDODunzps8AAAABLIAHIg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8900", "html_url": "https://github.com/huggingface/datasets/pull/8900", "diff_url": "https://github.com/huggingface/datasets/pull/8900.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8900.patch", "merged_at": null} | value:2026-10-27T21:41:44Z | value:https://api.github.com/repos/huggingface/datasets/issues/8900/labels{/name} | value:8900 | json:{"login": "Linxiushen", "id": 142557582, "node_id": "U_kgDOCH9Bjg", "avatar_url": "https://avatars.githubusercontent.com/u/142557582?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Linxiushen", "html_url": "https://github.com/Linxiushen", "followers_url": "https://api.github.com/users/Linxiushen/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8899/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8899 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8899/timeline | json:[] | value:6062954883 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8899 | value:fix(json): handle literal '/' in field names in mixed-type retry | value:2026-10-27T16:03:22Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8899/events | value:False | value:Fixes #8898
## Bug
When loading JSONL with the default `on_mixed_types="use_json"`, a record with a literal key such as `"a/b"` is misinterpreted: the PyArrow error column name is split into the nested path `["a", "b"]`, even though the record has one literal key `"a/b"`. Re-encoding with that nested path leaves... | value:https://api.github.com/repos/huggingface/datasets/issues/8899/comments | value:PR_kwDODunzps8AAAABLDweqg | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8899", "html_url": "https://github.com/huggingface/datasets/pull/8899", "diff_url": "https://github.com/huggingface/datasets/pull/8899.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8899.patch", "merged_at": null} | value:2026-11-03T16:51:54Z | value:https://api.github.com/repos/huggingface/datasets/issues/8899/labels{/name} | value:8899 | json:{"login": "Sailaja-ASB", "id": 83068760, "node_id": "MDQ6VXNlcjgzMDY4NzYw", "avatar_url": "https://avatars.githubusercontent.com/u/83068760?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Sailaja-ASB", "html_url": "https://github.com/Sailaja-ASB", "followers_url": "https://api.github.com/users/Sailaj... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8898/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | json:{"login": "its-amann", "id": 163747435, "node_id": "U_kgDOCcKWaw", "avatar_url": "https://avatars.githubusercontent.com/u/163747435?v=4", "gravatar_id": "", "url": "https://api.github.com/users/its-amann", "html_url": "https://github.com/its-amann", "followers_url": "https://api.github.com/users/its-amann/follower... | value:https://github.com/huggingface/datasets/issues/8898 | json:[{"login": "its-amann", "id": 163747435, "node_id": "U_kgDOCcKWaw", "avatar_url": "https://avatars.githubusercontent.com/u/163747435?v=4", "gravatar_id": "", "url": "https://api.github.com/users/its-amann", "html_url": "https://github.com/its-amann", "followers_url": "https://api.github.com/users/its-amann/followe... | value:https://api.github.com/repos/huggingface/datasets/issues/8898/timeline | json:[] | value:6054912822 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8898 | value:JSON loading retries indefinitely when a mixed-type field name contains a slash | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-27T07:34:57Z | value:completed | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8898/events | value:False | value:2026-11-09T17:28:52Z | value:### Describe the bug
On main `ef61228a5756dcebfff8ee245d89c72822fb6ae6` (`datasets` 5.2.0.dev0), loading a two-row JSONL file with a literal `a/b` field mixing a number and a string does not finish with the default `on_mixed_types="use_json"`.
The same values under a plain field name load as `Json()`. A slash-c... | value:https://api.github.com/repos/huggingface/datasets/issues/8898/comments | value:I_kwDODunzps8AAAABaOajNg | value:closed | value:2 | value:2026-11-09T17:28:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8898/labels{/name} | value:8898 | json:{"login": "bbbbbbbbbbb114", "id": 180641421, "node_id": "U_kgDOCsRejQ", "avatar_url": "https://avatars.githubusercontent.com/u/180641421?v=4", "gravatar_id": "", "url": "https://api.github.com/users/bbbbbbbbbbb114", "html_url": "https://github.com/bbbbbbbbbbb114", "followers_url": "https://api.github.com/users/bbb... | true | |||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8897/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8897 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8897/timeline | json:[] | value:6030725997 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8897 | value:Invalidate extraction caches when source archive metadata changes | value:2026-10-25T23:43:53Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8897/events | value:False | value:Replacing a local `.tar.gz` archive at the same path can return old rows from `load_dataset("json", ...)`, even after the builder detects the changed source and creates a new dataset fingerprint. The extraction cache currently identifies the archive by pathname alone.
This change includes the absolute path, file... | value:https://api.github.com/repos/huggingface/datasets/issues/8897/comments | value:PR_kwDODunzps8AAAABKpyD0A | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8897", "html_url": "https://github.com/huggingface/datasets/pull/8897", "diff_url": "https://github.com/huggingface/datasets/pull/8897.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8897.patch", "merged_at": null} | value:2026-11-03T15:10:07Z | value:https://api.github.com/repos/huggingface/datasets/issues/8897/labels{/name} | value:8897 | json:{"login": "omarborish", "id": 143956148, "node_id": "U_kgDOCJSYtA", "avatar_url": "https://avatars.githubusercontent.com/u/143956148?v=4", "gravatar_id": "", "url": "https://api.github.com/users/omarborish", "html_url": "https://github.com/omarborish", "followers_url": "https://api.github.com/users/omarborish/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8896/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8896 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8896/timeline | json:[] | value:6028268797 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8896 | value:Fix batched map on a dataset with an index when the function returns a table | value:2026-10-25T19:16:31Z | value:CONTRIBUTOR | json:{"login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers"... | value:https://api.github.com/repos/huggingface/datasets/issues/8896/events | value:False | value:2026-11-04T17:57:45Z | value:`cast()` and `cast_column()` fail on any dataset that has a FAISS or Elasticsearch index attached:
```python
import numpy as np
from datasets import Dataset, Value
ds = Dataset.from_dict({"id": list(range(5)), "v": np.eye(5, dtype=np.float32).tolist()})
ds.add_faiss_index("v")
ds.cast_column("id", Value("int32"... | value:https://api.github.com/repos/huggingface/datasets/issues/8896/comments | value:PR_kwDODunzps8AAAABKn2lxw | value:closed | value:2 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8896", "html_url": "https://github.com/huggingface/datasets/pull/8896", "diff_url": "https://github.com/huggingface/datasets/pull/8896.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8896.patch", "merged_at": "2026-11-04T17:57:45Z"} | value:2026-11-04T17:57:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8896/labels{/name} | value:8896 | json:{"login": "ViderStar", "id": 95582650, "node_id": "U_kgDOBbJ5ug", "avatar_url": "https://avatars.githubusercontent.com/u/95582650?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ViderStar", "html_url": "https://github.com/ViderStar", "followers_url": "https://api.github.com/users/ViderStar/followers"... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8895/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8895 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8895/timeline | json:[] | value:6027361872 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8895 | value:feat(features): add AtomicStructure with lazy AtomWorks decoding and preserved source files | value:2026-10-25T17:42:56Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8895/events | value:False | value:Adds `AtomicStructure` alongside `BioStructure` and `BioSequence`, preserving source files and loading atoms lazily with AtomWorks. PDB/mmCIF folder loaders use the new feature; existing serialized legacy feature names and sequence-loader schemas remain supported.
### Behavior
- Paths and bytes are stored uncha... | value:https://api.github.com/repos/huggingface/datasets/issues/8895/comments | value:PR_kwDODunzps8AAAABKnJDsQ | value:open | value:6 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8895", "html_url": "https://github.com/huggingface/datasets/pull/8895", "diff_url": "https://github.com/huggingface/datasets/pull/8895.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8895.patch", "merged_at": null} | value:2026-11-09T18:47:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8895/labels{/name} | value:8895 | json:{"login": "kierandidi", "id": 58345129, "node_id": "MDQ6VXNlcjU4MzQ1MTI5", "avatar_url": "https://avatars.githubusercontent.com/u/58345129?v=4", "gravatar_id": "", "url": "https://api.github.com/users/kierandidi", "html_url": "https://github.com/kierandidi", "followers_url": "https://api.github.com/users/kierandid... | value:True | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8894/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8894 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8894/timeline | json:[] | value:6025528770 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8894 | value:docs: propose evidence-first evolution intake and licensing gates | value:2026-10-25T14:28:37Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8894/events | value:False | value:https://api.github.com/repos/huggingface/datasets/issues/8894/comments | value:PR_kwDODunzps8AAAABKluSSg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8894", "html_url": "https://github.com/huggingface/datasets/pull/8894", "diff_url": "https://github.com/huggingface/datasets/pull/8894.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8894.patch", "merged_at": null} | value:2026-10-25T14:28:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8894/labels{/name} | value:8894 | json:{"login": "fisallllll280-code", "id": 259728808, "node_id": "U_kgDOD3slqA", "avatar_url": "https://avatars.githubusercontent.com/u/259728808?v=4", "gravatar_id": "", "url": "https://api.github.com/users/fisallllll280-code", "html_url": "https://github.com/fisallllll280-code", "followers_url": "https://api.github.c... | value:False | true | ||||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8893/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8893 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8893/timeline | json:[] | value:6023672883 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8893 | value:Fix xgetsize fallback to count bytes instead of decoded characters | value:2026-10-25T10:59:13Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8893/events | value:False | value:When the filesystem cannot report a file size, `xgetsize` measures the file by reading it, but it opens it with `xopen(path)`, whose default mode is `"r"`. So the fallback decodes the file as text: binary files raise `UnicodeDecodeError`, and text files return a character count after newline translation instead o... | value:https://api.github.com/repos/huggingface/datasets/issues/8893/comments | value:PR_kwDODunzps8AAAABKkSUqw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8893", "html_url": "https://github.com/huggingface/datasets/pull/8893", "diff_url": "https://github.com/huggingface/datasets/pull/8893.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8893.patch", "merged_at": null} | value:2026-10-25T11:13:59Z | value:https://api.github.com/repos/huggingface/datasets/issues/8893/labels{/name} | value:8893 | json:{"login": "Sonica-B", "id": 66372798, "node_id": "MDQ6VXNlcjY2MzcyNzk4", "avatar_url": "https://avatars.githubusercontent.com/u/66372798?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Sonica-B", "html_url": "https://github.com/Sonica-B", "followers_url": "https://api.github.com/users/Sonica-B/follow... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8892/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8892 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8892/timeline | json:[] | value:6018163293 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8892 | value:Forward ignore_decode_attribute in nested require_decoding calls | value:2026-10-24T22:12:46Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8892/events | value:False | value:## What
`require_decoding(feature, ignore_decode_attribute=...)` drops `ignore_decode_attribute` when it recurses: the `dict`, `list`/`tuple`, `LargeList` and `List` branches call `require_decoding(f)` without forwarding the flag. Only a top-level feature honours it.
`ignore_decode_attribute=True` is passed by ... | value:https://api.github.com/repos/huggingface/datasets/issues/8892/comments | value:PR_kwDODunzps8AAAABKf_y5A | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8892", "html_url": "https://github.com/huggingface/datasets/pull/8892", "diff_url": "https://github.com/huggingface/datasets/pull/8892.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8892.patch", "merged_at": null} | value:2026-10-24T22:12:46Z | value:https://api.github.com/repos/huggingface/datasets/issues/8892/labels{/name} | value:8892 | json:{"login": "udsy19", "id": 71968569, "node_id": "MDQ6VXNlcjcxOTY4NTY5", "avatar_url": "https://avatars.githubusercontent.com/u/71968569?v=4", "gravatar_id": "", "url": "https://api.github.com/users/udsy19", "html_url": "https://github.com/udsy19", "followers_url": "https://api.github.com/users/udsy19/followers", "f... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8891/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8891 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8891/timeline | json:[] | value:6014808547 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8891 | value:Don't infer the Harbor builder from unrelated .md and .toml files | value:2026-10-24T15:58:23Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8891/events | value:False | value:Since #8729, `.md` and `.toml` are allowed data file extensions. Every Markdown or TOML file matched by the default `**` pattern now counts as a vote for `harbor` when the builder is inferred. So a directory with `data.jsonl`, `LICENSE.md` and `CHANGELOG.md` is inferred as Harbor, and so is `data.csv` with `pypro... | value:https://api.github.com/repos/huggingface/datasets/issues/8891/comments | value:PR_kwDODunzps8AAAABKdZ_aw | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8891", "html_url": "https://github.com/huggingface/datasets/pull/8891", "diff_url": "https://github.com/huggingface/datasets/pull/8891.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8891.patch", "merged_at": null} | value:2026-10-24T15:58:23Z | value:https://api.github.com/repos/huggingface/datasets/issues/8891/labels{/name} | value:8891 | json:{"login": "NgoQuocViet2001", "id": 123613986, "node_id": "U_kgDOB14zIg", "avatar_url": "https://avatars.githubusercontent.com/u/123613986?v=4", "gravatar_id": "", "url": "https://api.github.com/users/NgoQuocViet2001", "html_url": "https://github.com/NgoQuocViet2001", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8890/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8890 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8890/timeline | json:[] | value:6014230128 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8890 | value:keep nullable integers intact in to_pandas | value:2026-10-24T14:54:55Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8890/events | value:False | value:an int64 column with a null was converted to float64. float64 cannot hold every int64, so `2**62 + 1` came back as a different number. `to_pandas` and the pandas formatter now keep those nulls as missing values, the same way `to_csv` / `to_json` / `to_sql` already do. a column with no nulls stays int64.
checked ... | value:https://api.github.com/repos/huggingface/datasets/issues/8890/comments | value:PR_kwDODunzps8AAAABKc91DA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8890", "html_url": "https://github.com/huggingface/datasets/pull/8890", "diff_url": "https://github.com/huggingface/datasets/pull/8890.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8890.patch", "merged_at": null} | value:2026-10-24T14:54:55Z | value:https://api.github.com/repos/huggingface/datasets/issues/8890/labels{/name} | value:8890 | json:{"login": "chessing234", "id": 134995372, "node_id": "U_kgDOCAvdrA", "avatar_url": "https://avatars.githubusercontent.com/u/134995372?v=4", "gravatar_id": "", "url": "https://api.github.com/users/chessing234", "html_url": "https://github.com/chessing234", "followers_url": "https://api.github.com/users/chessing234/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8889/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8889 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8889/timeline | json:[] | value:6008187639 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8889 | value:Keep row validity and payloads in fixed-shape ArrayXD pandas round trips | value:2026-10-24T01:50:23Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8889/events | value:False | value:Fixes #8888
`PandasArrayExtensionDtype.__from_arrow__` materialized every fixed-shape ArrayXD column through `ArrayExtensionArray.to_numpy`, which pads the null rows with NaN and promotes the whole column to float64. Two consequences at the pandas boundary: the Arrow row validity was gone (a null row and a legit... | value:https://api.github.com/repos/huggingface/datasets/issues/8889/comments | value:PR_kwDODunzps8AAAABKYT83w | value:open | value:5 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8889", "html_url": "https://github.com/huggingface/datasets/pull/8889", "diff_url": "https://github.com/huggingface/datasets/pull/8889.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8889.patch", "merged_at": null} | value:2026-11-02T04:14:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8889/labels{/name} | value:8889 | json:{"login": "choksi2212", "id": 162600900, "node_id": "U_kgDOCbEXxA", "avatar_url": "https://avatars.githubusercontent.com/u/162600900?v=4", "gravatar_id": "", "url": "https://api.github.com/users/choksi2212", "html_url": "https://github.com/choksi2212", "followers_url": "https://api.github.com/users/choksi2212/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8888/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/issues/8888 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8888/timeline | json:[] | value:6006545971 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8888 | value:Fixed-shape ArrayXD pandas round trips lose row nulls and integer precision | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-10-23T22:45:51Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8888/events | value:False | value:### Describe the bug
`Dataset.from_pandas(ds.to_pandas())` loses missing rows in a fixed-shape ArrayXD column. A null row becomes a valid all-NaN array. A legitimate all-NaN array must remain distinct from that missing row.
### Steps to reproduce the bug
Reproduced on main `ef61228a5756dcebfff8ee245d89c72822fb... | value:https://api.github.com/repos/huggingface/datasets/issues/8888/comments | value:I_kwDODunzps8AAAABZgSeMw | value:open | value:2 | value:2026-10-24T01:50:42Z | value:https://api.github.com/repos/huggingface/datasets/issues/8888/labels{/name} | value:8888 | json:{"login": "sjdoane", "id": 234531224, "node_id": "U_kgDODfqpmA", "avatar_url": "https://avatars.githubusercontent.com/u/234531224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sjdoane", "html_url": "https://github.com/sjdoane", "followers_url": "https://api.github.com/users/sjdoane/followers", "fol... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8887/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8887 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8887/timeline | json:[] | value:6001922606 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8887 | value:Preserve selected streaming rows across concatenated epochs | value:2026-10-23T16:51:29Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8887/events | value:False | value:After concatenating a streaming dataset selected with `take(4)`, `set_epoch(1)` can change its rows from `[0, 1, 2, 3]` to `[0, 3, 6, 9]`. The same problem affects `skip()` and explicit shuffling after concatenation.
Propagate shuffling through vertical concatenation's shard wrappers so the existing skip/take or... | value:https://api.github.com/repos/huggingface/datasets/issues/8887/comments | value:PR_kwDODunzps8AAAABKTRcRg | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8887", "html_url": "https://github.com/huggingface/datasets/pull/8887", "diff_url": "https://github.com/huggingface/datasets/pull/8887.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8887.patch", "merged_at": null} | value:2026-10-23T16:51:29Z | value:https://api.github.com/repos/huggingface/datasets/issues/8887/labels{/name} | value:8887 | json:{"login": "sjdoane", "id": 234531224, "node_id": "U_kgDODfqpmA", "avatar_url": "https://avatars.githubusercontent.com/u/234531224?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sjdoane", "html_url": "https://github.com/sjdoane", "followers_url": "https://api.github.com/users/sjdoane/followers", "fol... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8886/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8886 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8886/timeline | json:[] | value:5996193964 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8886 | value:Fix SQLAlchemy query cache collisions for bound parameters | value:2026-10-23T10:33:03Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8886/events | value:False | value:
## What does this PR do?
Fixes #8877
`Dataset.from_sql()` can silently return incorrect cached rows when two SQLAlchemy queries differ only in their bound parameter values. The existing cache identity includes the compiled SQL placeholders but not their values, causing queries with different `WHERE`, `LIMIT/OF... | value:https://api.github.com/repos/huggingface/datasets/issues/8886/comments | value:PR_kwDODunzps8AAAABKOpggg | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8886", "html_url": "https://github.com/huggingface/datasets/pull/8886", "diff_url": "https://github.com/huggingface/datasets/pull/8886.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8886.patch", "merged_at": null} | value:2026-10-25T10:48:31Z | value:https://api.github.com/repos/huggingface/datasets/issues/8886/labels{/name} | value:8886 | json:{"login": "gorkadabo1", "id": 106764709, "node_id": "U_kgDOBl0ZpQ", "avatar_url": "https://avatars.githubusercontent.com/u/106764709?v=4", "gravatar_id": "", "url": "https://api.github.com/users/gorkadabo1", "html_url": "https://github.com/gorkadabo1", "followers_url": "https://api.github.com/users/gorkadabo1/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8885/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8885 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8885/timeline | json:[] | value:5996029451 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8885 | value:Close Parquet export writers with a context manager | value:2026-10-23T10:20:18Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8885/events | value:False | value:Fixes #8743.
Use `pyarrow.parquet.ParquetWriter` as a context manager so query, batch-write, and metadata failures finalize the writer before the enclosing path-owned file handle closes. Caller-provided binary streams remain open, and the successful byte-count result is unchanged.
This is an alternative to #874... | value:https://api.github.com/repos/huggingface/datasets/issues/8885/comments | value:PR_kwDODunzps8AAAABKOhAgA | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8885", "html_url": "https://github.com/huggingface/datasets/pull/8885", "diff_url": "https://github.com/huggingface/datasets/pull/8885.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8885.patch", "merged_at": null} | value:2026-10-23T10:20:18Z | value:https://api.github.com/repos/huggingface/datasets/issues/8885/labels{/name} | value:8885 | json:{"login": "wpyaoxueros", "id": 139048865, "node_id": "U_kgDOCEm3oQ", "avatar_url": "https://avatars.githubusercontent.com/u/139048865?v=4", "gravatar_id": "", "url": "https://api.github.com/users/wpyaoxueros", "html_url": "https://github.com/wpyaoxueros", "followers_url": "https://api.github.com/users/wpyaoxueros/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8884/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8884 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8884/timeline | json:[] | value:5995468709 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8884 | value:Skip the spark tests when pyspark is not installed | value:2026-10-23T09:38:47Z | value:NONE | json:{"login": "ege-arhan", "id": 68082325, "node_id": "MDQ6VXNlcjY4MDgyMzI1", "avatar_url": "https://avatars.githubusercontent.com/u/68082325?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ege-arhan", "html_url": "https://github.com/ege-arhan", "followers_url": "https://api.github.com/users/ege-arhan/fo... | value:https://api.github.com/repos/huggingface/datasets/issues/8884/events | value:False | value:2026-10-24T23:28:21Z | value:Running the test suite without the optional `pyspark` dependency aborts collection instead of skipping the spark tests:
```
ERROR tests/packaged_modules/test_spark.py
ModuleNotFoundError: No module named 'pyspark'
```
`tests/packaged_modules/test_spark.py` imports `pyspark` at module level, so collection fails ... | value:https://api.github.com/repos/huggingface/datasets/issues/8884/comments | value:PR_kwDODunzps8AAAABKOEJ2g | value:closed | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8884", "html_url": "https://github.com/huggingface/datasets/pull/8884", "diff_url": "https://github.com/huggingface/datasets/pull/8884.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8884.patch", "merged_at": null} | value:2026-10-24T23:28:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8884/labels{/name} | value:8884 | json:{"login": "ege-arhan", "id": 68082325, "node_id": "MDQ6VXNlcjY4MDgyMzI1", "avatar_url": "https://avatars.githubusercontent.com/u/68082325?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ege-arhan", "html_url": "https://github.com/ege-arhan", "followers_url": "https://api.github.com/users/ege-arhan/fo... | value:False | true |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.