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/8475/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/8475 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8475/timeline | json:[] | value:5137335145 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8475 | value:`Dataset.select` raises a raw OverflowError on in-range negative indices that its own validation allows | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-13T03:02:44Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8475/events | value:False | value:### Describe the bug
`Dataset.select` validates indices with [`_check_valid_indices_value`](https://github.com/huggingface/datasets/blob/836b82e0544cabf6474b25ade131b4d21e570373/src/datasets/arrow_dataset.py#L636-L638), which implements Python-style negative indexing: it only raises for a negative index when `in... | value:https://api.github.com/repos/huggingface/datasets/issues/8475/comments | value:I_kwDODunzps8AAAABMjWDaQ | value:open | value:2 | value:2026-08-16T23:36:56Z | value:https://api.github.com/repos/huggingface/datasets/issues/8475/labels{/name} | value:8475 | json:{"login": "codeAnqiang-ma", "id": 273298913, "node_id": "U_kgDOEEo14Q", "avatar_url": "https://avatars.githubusercontent.com/u/273298913?v=4", "gravatar_id": "", "url": "https://api.github.com/users/codeAnqiang-ma", "html_url": "https://github.com/codeAnqiang-ma", "followers_url": "https://api.github.com/users/cod... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8474/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/8474 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8474/timeline | json:[] | value:5137329748 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8474 | value:Fix negative percent split boundary anchoring to the start when it rounds to 0 | value:2026-08-13T03:01:41Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8474/events | value:False | value:Fixes #8473.
### What this fixes
With the default `closest` rounding, a negative percent split boundary whose absolute value rounds to 0 loses its sign, so "relative to the end" silently becomes "relative to the start":
```python
# on a 30-example split
train[-1%:] -> (0, 30) all 30 rows, while train[99%:]... | value:https://api.github.com/repos/huggingface/datasets/issues/8474/comments | value:PR_kwDODunzps7-X9tn | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8474", "html_url": "https://github.com/huggingface/datasets/pull/8474", "diff_url": "https://github.com/huggingface/datasets/pull/8474.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8474.patch", "merged_at": null} | value:2026-08-13T03:01:41Z | value:https://api.github.com/repos/huggingface/datasets/issues/8474/labels{/name} | value:8474 | json:{"login": "codeAnqiang-ma", "id": 273298913, "node_id": "U_kgDOEEo14Q", "avatar_url": "https://avatars.githubusercontent.com/u/273298913?v=4", "gravatar_id": "", "url": "https://api.github.com/users/codeAnqiang-ma", "html_url": "https://github.com/codeAnqiang-ma", "followers_url": "https://api.github.com/users/cod... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8473/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/8473 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8473/timeline | json:[] | value:5137305743 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8473 | value:Negative percent split boundary silently anchors to the start when it rounds to 0 (`train[-1%:]` returns the whole split) | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-13T02:56:52Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8473/events | value:False | value:### Describe the bug
With the default `closest` rounding, a negative percent split boundary whose absolute value rounds to 0 loses its sign, so "relative to the end" silently becomes "relative to the start":
- `train[-1%:]` on a 30-example split returns **all 30 rows** instead of the last 1% (~0 rows, as `train... | value:https://api.github.com/repos/huggingface/datasets/issues/8473/comments | value:I_kwDODunzps8AAAABMjUQjw | value:open | value:0 | value:2026-08-13T02:56:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8473/labels{/name} | value:8473 | json:{"login": "codeAnqiang-ma", "id": 273298913, "node_id": "U_kgDOEEo14Q", "avatar_url": "https://avatars.githubusercontent.com/u/273298913?v=4", "gravatar_id": "", "url": "https://api.github.com/users/codeAnqiang-ma", "html_url": "https://github.com/codeAnqiang-ma", "followers_url": "https://api.github.com/users/cod... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8472/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/8472 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8472/timeline | json:[] | value:5128543479 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8472 | value:batch(0) silently returns one batch with the whole dataset instead of raising | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-12T07:52:28Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8472/events | value:False | value:### Describe the bug
`Dataset.batch()` and `IterableDataset.batch()` accept `batch_size=0` and negative sizes, and quietly return **one batch holding the whole dataset**. Asking for a zero-sized batch produces the largest possible batch, which is as close to the opposite of the request as the API can get.
### S... | value:https://api.github.com/repos/huggingface/datasets/issues/8472/comments | value:I_kwDODunzps8AAAABMa9c9w | value:open | value:2 | value:2026-08-16T15:56:35Z | value:https://api.github.com/repos/huggingface/datasets/issues/8472/labels{/name} | value:8472 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8471/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/8471 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8471/timeline | json:[] | value:5128533318 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8471 | value:Stop `Dataset.sort()` emitting a pyarrow FutureWarning | value:2026-08-12T07:51:09Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8471/events | value:False | value:## The bug
Every `Dataset.sort()` call emits a pyarrow `FutureWarning` on pyarrow >= 25:
```
FutureWarning: Specifying null_placement in SortOptions is deprecated as of 25.0.0.
Specify null_placement per sort_key instead.
```
It comes from `Dataset.sort` itself, not from user code:
```python
indices = pc.sort... | value:https://api.github.com/repos/huggingface/datasets/issues/8471/comments | value:PR_kwDODunzps797k6o | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8471", "html_url": "https://github.com/huggingface/datasets/pull/8471", "diff_url": "https://github.com/huggingface/datasets/pull/8471.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8471.patch", "merged_at": null} | value:2026-08-12T07:57:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8471/labels{/name} | value:8471 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8470/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/8470 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8470/timeline | json:[] | value:5128425500 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8470 | value:Keep the dtype of an empty column in the numpy formatter | value:2026-08-12T07:38:03Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8470/events | value:False | value:Fixes #8469.
## The bug
Under `numpy` format, a column holding zero rows comes back as `float32` regardless of its declared dtype — while the same column keeps its dtype as soon as it holds one row, and `pandas`/`arrow` keep it either way.
```python
feats = Features({"a": Value("int64")})
ds = Dataset.from_dic... | value:https://api.github.com/repos/huggingface/datasets/issues/8470/comments | value:PR_kwDODunzps797OZ4 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8470", "html_url": "https://github.com/huggingface/datasets/pull/8470", "diff_url": "https://github.com/huggingface/datasets/pull/8470.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8470.patch", "merged_at": null} | value:2026-08-12T07:38:03Z | value:https://api.github.com/repos/huggingface/datasets/issues/8470/labels{/name} | value:8470 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8469/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/8469 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8469/timeline | json:[] | value:5128417961 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8469 | value:Empty column loses its dtype under numpy format (always float32) | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-12T07:37:13Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8469/events | value:False | value:### Describe the bug
Under `numpy` format, a column that holds zero rows comes back as `float32` no matter what dtype it was declared with. The same column keeps its dtype as soon as it holds a single row, and the `pandas` and `arrow` formatters keep it in both cases.
This is easy to hit in normal code: any `fi... | value:https://api.github.com/repos/huggingface/datasets/issues/8469/comments | value:I_kwDODunzps8AAAABMa1yqQ | value:open | value:0 | value:2026-08-12T07:38:05Z | value:https://api.github.com/repos/huggingface/datasets/issues/8469/labels{/name} | value:8469 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8468/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/8468 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8468/timeline | json:[] | value:5128312016 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8468 | value:Implement `construct_from_string` on `PandasArrayExtensionDtype` | value:2026-08-12T07:25:39Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8468/events | value:False | value:Fixes #8467.
## The bug
`PandasArrayExtensionDtype` never implements `construct_from_string`, so it inherits the base `ExtensionDtype` version:
```python
@classmethod
def construct_from_string(cls, string):
assert isinstance(cls.name, str), (cls, type(cls.name))
```
`name` is an instance `@property` here,... | value:https://api.github.com/repos/huggingface/datasets/issues/8468/comments | value:PR_kwDODunzps7962H- | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8468", "html_url": "https://github.com/huggingface/datasets/pull/8468", "diff_url": "https://github.com/huggingface/datasets/pull/8468.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8468.patch", "merged_at": null} | value:2026-08-12T07:25:39Z | value:https://api.github.com/repos/huggingface/datasets/issues/8468/labels{/name} | value:8468 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8467/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/8467 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8467/timeline | json:[] | value:5128306077 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8467 | value:PandasArrayExtensionDtype has no construct_from_string, so comparing the dtype to any string raises AssertionError | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-12T07:24:51Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8467/events | value:False | value:### Describe the bug
`PandasArrayExtensionDtype` does not implement `construct_from_string`, so it inherits the base `ExtensionDtype` version, which opens with:
```python
@classmethod
def construct_from_string(cls, string):
assert isinstance(cls.name, str), (cls, type(cls.name))
```
`name` is an instance `... | value:https://api.github.com/repos/huggingface/datasets/issues/8467/comments | value:I_kwDODunzps8AAAABMau9nQ | value:open | value:0 | value:2026-08-12T07:25:58Z | value:https://api.github.com/repos/huggingface/datasets/issues/8467/labels{/name} | value:8467 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8466/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/8466 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8466/timeline | json:[] | value:5128026958 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8466 | value:Raise a clear error for a deeply nested mesh file | value:2026-08-12T06:46:53Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8466/events | value:False | value:Fixes #8246.
## The bug
A GLB file carries its scene description as an embedded JSON chunk, and that JSON can be nested arbitrarily deeply. Parsing it exhausts the interpreter stack, so `Mesh.decode_example` propagates a bare
```
RecursionError: maximum recursion depth exceeded while decoding a JSON object fro... | value:https://api.github.com/repos/huggingface/datasets/issues/8466/comments | value:PR_kwDODunzps7957wa | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8466", "html_url": "https://github.com/huggingface/datasets/pull/8466", "diff_url": "https://github.com/huggingface/datasets/pull/8466.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8466.patch", "merged_at": null} | value:2026-08-12T06:46:53Z | value:https://api.github.com/repos/huggingface/datasets/issues/8466/labels{/name} | value:8466 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8465/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/8465 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8465/timeline | json:[] | value:5127966298 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8465 | value:PandasArrayExtensionArray.take coerces fill_value before checking whether anything is filled, breaking boolean masks on integer array columns | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-12T06:38:18Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8465/events | value:False | value:### Describe the bug
`PandasArrayExtensionArray.take` coerces `fill_value` to the array's value type **before** checking whether any element will actually be filled:
https://github.com/huggingface/datasets/blob/main/src/datasets/features/features.py#L965-L972
```python
if allow_fill:
fill_value = (
... | value:https://api.github.com/repos/huggingface/datasets/issues/8465/comments | value:I_kwDODunzps8AAAABMaaOWg | value:open | value:0 | value:2026-08-12T06:38:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8465/labels{/name} | value:8465 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8464/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/8464 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8464/timeline | json:[] | value:5127916520 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8464 | value:Make `PandasArrayExtensionDtype._metadata` a tuple | value:2026-08-12T06:30:49Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8464/events | value:False | value:Fixes #8375.
## The bug
`PandasArrayExtensionDtype._metadata` was the plain string `"value_type"`. Pandas expects a **tuple of attribute names** and iterates over it in `ExtensionDtype.__eq__` and `ExtensionDtype.__hash__`:
```python
all(getattr(self, attr) == getattr(other, attr) for attr in self._metadata)
`... | value:https://api.github.com/repos/huggingface/datasets/issues/8464/comments | value:PR_kwDODunzps795lBw | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8464", "html_url": "https://github.com/huggingface/datasets/pull/8464", "diff_url": "https://github.com/huggingface/datasets/pull/8464.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8464.patch", "merged_at": null} | value:2026-08-23T02:13:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8464/labels{/name} | value:8464 | json:{"login": "shashvat-singham", "id": 78735607, "node_id": "MDQ6VXNlcjc4NzM1NjA3", "avatar_url": "https://avatars.githubusercontent.com/u/78735607?v=4", "gravatar_id": "", "url": "https://api.github.com/users/shashvat-singham", "html_url": "https://github.com/shashvat-singham", "followers_url": "https://api.github.c... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8463/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/8463 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8463/timeline | json:[] | value:5126995924 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8463 | value:very bad quality images in most of them | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-12T03:49:23Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8463/events | value:False | value:### Describe the bug
very bad quality images in most of them
### Steps to reproduce the bug
.
### Expected behavior
.
### Environment info
. | value:https://api.github.com/repos/huggingface/datasets/issues/8463/comments | value:I_kwDODunzps8AAAABMZe_1A | value:open | value:0 | value:2026-08-12T03:49:23Z | value:https://api.github.com/repos/huggingface/datasets/issues/8463/labels{/name} | value:8463 | json:{"login": "anis-mselmi", "id": 209621762, "node_id": "U_kgDODH6TAg", "avatar_url": "https://avatars.githubusercontent.com/u/209621762?v=4", "gravatar_id": "", "url": "https://api.github.com/users/anis-mselmi", "html_url": "https://github.com/anis-mselmi", "followers_url": "https://api.github.com/users/anis-mselmi/... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8462/reactions", "total_count": 1, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 1, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8462 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8462/timeline | json:[] | value:5121965559 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8462 | value:Fix info from data_dir/data_files or malformed yaml | value:2026-08-11T15:45:50Z | 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/8462/events | value:False | value:2026-08-12T14:56:49Z | value:- don't reuse the info from a config if data_files or data_dir are passed
- ignore malformed split info in yaml
cc @anton-l | value:https://api.github.com/repos/huggingface/datasets/issues/8462/comments | value:PR_kwDODunzps79mZju | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8462", "html_url": "https://github.com/huggingface/datasets/pull/8462", "diff_url": "https://github.com/huggingface/datasets/pull/8462.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8462.patch", "merged_at": "2026-08-12T14:56:49Z"} | value:2026-08-12T14:56:52Z | value:https://api.github.com/repos/huggingface/datasets/issues/8462/labels{/name} | value:8462 | 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/8461/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/8461 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8461/timeline | json:[] | value:5121551071 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8461 | value:Add test for tokenizer stride with Dataset.map (issue #4769) | value:2026-08-11T15:00:25Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8461/events | value:False | value:## Problem
When processing SQuAD v1.1 datasets with `max_seq_length=128` and `doc_stride=96` using transformers' QA pipeline, users encountered an assertion failure:
```
thread '<unnamed>' panicked at 'assertion failed: stride < max_len'
```
This prevented legitimate use of the stride parameter in question-answe... | value:https://api.github.com/repos/huggingface/datasets/issues/8461/comments | value:PR_kwDODunzps79lDrU | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8461", "html_url": "https://github.com/huggingface/datasets/pull/8461", "diff_url": "https://github.com/huggingface/datasets/pull/8461.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8461.patch", "merged_at": null} | value:2026-08-13T15:02:47Z | value:https://api.github.com/repos/huggingface/datasets/issues/8461/labels{/name} | value:8461 | json:{"login": "imjimit07", "id": 78078749, "node_id": "MDQ6VXNlcjc4MDc4NzQ5", "avatar_url": "https://avatars.githubusercontent.com/u/78078749?v=4", "gravatar_id": "", "url": "https://api.github.com/users/imjimit07", "html_url": "https://github.com/imjimit07", "followers_url": "https://api.github.com/users/imjimit07/fo... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8460/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/8460 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8460/timeline | json:[] | value:5117338188 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8460 | value:Fix Audio encoding for path rows with null arrays | value:2026-08-11T06:26:20Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8460/events | value:False | value:## Summary
- treat a null `array` value as absent in `Audio.encode_example()`, so path-backed rows reach the existing path branch
- cover the null-array case in the existing `encode_example` and `cast_to_audio_features` parametrize lists
## Problem
Casting a struct column to `Audio()` fails when a row is path-... | value:https://api.github.com/repos/huggingface/datasets/issues/8460/comments | value:PR_kwDODunzps79XWqT | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8460", "html_url": "https://github.com/huggingface/datasets/pull/8460", "diff_url": "https://github.com/huggingface/datasets/pull/8460.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8460.patch", "merged_at": null} | value:2026-08-14T07:24:03Z | value:https://api.github.com/repos/huggingface/datasets/issues/8460/labels{/name} | value:8460 | json:{"login": "Losera", "id": 141312508, "node_id": "U_kgDOCGxB_A", "avatar_url": "https://avatars.githubusercontent.com/u/141312508?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Losera", "html_url": "https://github.com/Losera", "followers_url": "https://api.github.com/users/Losera/followers", "followi... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8459/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/8459 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8459/timeline | json:[] | value:5115089650 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8459 | value:Fix Dataset.to_json() truncating/corrupting non-millisecond temporal columns | value:2026-08-10T23:07:21Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8459/events | value:False | value:Fixes #8390
## Problem
`Dataset.to_json()` writes every timestamp/duration column through `pandas.DataFrame.to_json()`, which only supports a single global `date_unit` for the entire call (defaulting to milliseconds).
This ignores each column's actual Arrow unit (`s`/`ms`/`us`/`ns`), causing a `to_json()... | value:https://api.github.com/repos/huggingface/datasets/issues/8459/comments | value:PR_kwDODunzps79QNXm | value:open | value:2 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8459", "html_url": "https://github.com/huggingface/datasets/pull/8459", "diff_url": "https://github.com/huggingface/datasets/pull/8459.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8459.patch", "merged_at": null} | value:2026-08-12T22:02:19Z | value:https://api.github.com/repos/huggingface/datasets/issues/8459/labels{/name} | value:8459 | json:{"login": "aakif-ms", "id": 144618319, "node_id": "U_kgDOCJ6zTw", "avatar_url": "https://avatars.githubusercontent.com/u/144618319?v=4", "gravatar_id": "", "url": "https://api.github.com/users/aakif-ms", "html_url": "https://github.com/aakif-ms", "followers_url": "https://api.github.com/users/aakif-ms/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8458/reactions", "total_count": 0, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[{"id": 1935892871, "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement", "name": "enhancement", "color": "a2eeef", "default": true, "description": "New feature or request", "archived_at": null, "archived_by": null}] | value:https://github.com/huggingface/datasets/issues/8458 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8458/timeline | json:[] | value:5110353136 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8458 | value:Replace httpx with httpx2 | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-10T13:01:44Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8458/events | value:False | value:### Feature request
Replace the httpx library with the [httpx2](https://github.com/pydantic/httpx2) fork, maintained by Pydantic.
### Motivation
Httpx has not been released since December 2024, and has closed all public access to Issues / Discussions / PRs, this raises questions about the actual state of maint... | value:https://api.github.com/repos/huggingface/datasets/issues/8458/comments | value:I_kwDODunzps8AAAABMJnM8A | value:open | value:3 | value:2026-08-16T15:56:33Z | value:https://api.github.com/repos/huggingface/datasets/issues/8458/labels{/name} | value:8458 | json:{"login": "Aegdesil", "id": 7816329, "node_id": "MDQ6VXNlcjc4MTYzMjk=", "avatar_url": "https://avatars.githubusercontent.com/u/7816329?v=4", "gravatar_id": "", "url": "https://api.github.com/users/Aegdesil", "html_url": "https://github.com/Aegdesil", "followers_url": "https://api.github.com/users/Aegdesil/follower... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8457/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/8457 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8457/timeline | json:[] | value:5105363162 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8457 | value:fix(types): expose PyTorch dataset compatibility | value:2026-08-09T23:58:43Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8457/events | value:False | value:## What does this PR do?
Make `Dataset` and `IterableDataset` visible to static type checkers as compatible with PyTorch's dataset classes, so Pyright accepts both map-style and iterable datasets as `torch.utils.data.DataLoader` inputs, including `dataset.with_format(torch)`. This addresses the typing gap descri... | value:https://api.github.com/repos/huggingface/datasets/issues/8457/comments | value:PR_kwDODunzps78w-i9 | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8457", "html_url": "https://github.com/huggingface/datasets/pull/8457", "diff_url": "https://github.com/huggingface/datasets/pull/8457.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8457.patch", "merged_at": null} | value:2026-08-14T23:43:27Z | value:https://api.github.com/repos/huggingface/datasets/issues/8457/labels{/name} | value:8457 | json:{"login": "uuzzrm", "id": 267482477, "node_id": "U_kgDOD_F1bQ", "avatar_url": "https://avatars.githubusercontent.com/u/267482477?v=4", "gravatar_id": "", "url": "https://api.github.com/users/uuzzrm", "html_url": "https://github.com/uuzzrm", "followers_url": "https://api.github.com/users/uuzzrm/followers", "followi... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8456/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/8456 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8456/timeline | json:[] | value:5103866382 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8456 | value:Fix Dataset.map(batched=True, drop_last_batch=True) returning unchanged input when dataset is smaller than batch_size | value:2026-08-09T17:43:50Z | value:NONE | json:{"login": "mayuriphad", "id": 163738104, "node_id": "U_kgDOCcJx-A", "avatar_url": "https://avatars.githubusercontent.com/u/163738104?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mayuriphad", "html_url": "https://github.com/mayuriphad", "followers_url": "https://api.github.com/users/mayuriphad/foll... | value:https://api.github.com/repos/huggingface/datasets/issues/8456/events | value:False | value:2026-09-18T08:30:19Z | value:Fixes #8386
## Bug
When `drop_last_batch=True` and the dataset has fewer rows than `batch_size`, `Dataset.map(batched=True, ...)` was returning the **input dataset unchanged** instead of an empty dataset. Every batch is incomplete in this situation, so the mapped function is never invoked and `update_data` stay... | value:https://api.github.com/repos/huggingface/datasets/issues/8456/comments | value:PR_kwDODunzps78sbms | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8456", "html_url": "https://github.com/huggingface/datasets/pull/8456", "diff_url": "https://github.com/huggingface/datasets/pull/8456.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8456.patch", "merged_at": null} | value:2026-09-18T08:30:19Z | value:https://api.github.com/repos/huggingface/datasets/issues/8456/labels{/name} | value:8456 | json:{"login": "mayuriphad", "id": 163738104, "node_id": "U_kgDOCcJx-A", "avatar_url": "https://avatars.githubusercontent.com/u/163738104?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mayuriphad", "html_url": "https://github.com/mayuriphad", "followers_url": "https://api.github.com/users/mayuriphad/foll... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8455/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/8455 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8455/timeline | json:[] | value:5101882207 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8455 | value:close input stream in _in_memory_arrow_table_from_file to prevent file descriptor leaks | value:2026-08-09T09:40:16Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8455/events | value:False | value:In `src/datasets/table.py`, `_in_memory_arrow_table_from_file` opens a `pa.input_stream(filename)` without a `with` context manager block.
When exceptions occur during stream parsing or when exception tracebacks retain frame variables in memory (e.g., in logging handlers or data loader error queues), local `in_... | value:https://api.github.com/repos/huggingface/datasets/issues/8455/comments | value:PR_kwDODunzps78mb3T | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8455", "html_url": "https://github.com/huggingface/datasets/pull/8455", "diff_url": "https://github.com/huggingface/datasets/pull/8455.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8455.patch", "merged_at": null} | value:2026-08-12T07:09:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8455/labels{/name} | value:8455 | json:{"login": "HassanRady", "id": 48225007, "node_id": "MDQ6VXNlcjQ4MjI1MDA3", "avatar_url": "https://avatars.githubusercontent.com/u/48225007?v=4", "gravatar_id": "", "url": "https://api.github.com/users/HassanRady", "html_url": "https://github.com/HassanRady", "followers_url": "https://api.github.com/users/HassanRad... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8454/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/8454 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8454/timeline | json:[] | value:5101287981 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8454 | value:Coerce ClassLabel.names to plain str to fix invalid YAML in README.md | value:2026-08-09T06:50:20Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8454/events | value:False | value:Coerce ClassLabel.names to plain str to fix invalid YAML in README.md
Fixes #6919
## Problem
`push_to_hub` can fail with:
```
ValueError: Invalid metadata in README.md.
- Invalid YAML in README.md: unknown tag !<tag:yaml.org,2002:python/tuple> (50:11)
```
This happens whenever a `ClassLabel`'s `names` list c... | value:https://api.github.com/repos/huggingface/datasets/issues/8454/comments | value:PR_kwDODunzps78kqeD | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8454", "html_url": "https://github.com/huggingface/datasets/pull/8454", "diff_url": "https://github.com/huggingface/datasets/pull/8454.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8454.patch", "merged_at": null} | value:2026-08-12T07:12:01Z | value:https://api.github.com/repos/huggingface/datasets/issues/8454/labels{/name} | value:8454 | json:{"login": "RudrenduPaul", "id": 38769913, "node_id": "MDQ6VXNlcjM4NzY5OTEz", "avatar_url": "https://avatars.githubusercontent.com/u/38769913?v=4", "gravatar_id": "", "url": "https://api.github.com/users/RudrenduPaul", "html_url": "https://github.com/RudrenduPaul", "followers_url": "https://api.github.com/users/Rud... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8453/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/8453 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8453/timeline | json:[] | value:5100509939 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8453 | value:Reject features with different columns in IterableDataset.cast | value:2026-08-09T02:25:09Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8453/events | value:False | value:`IterableDataset.cast()` assigns the given features directly:
```python
features = _fix_for_backward_compatible_features(features)
info = self._info.copy()
info.features = features
```
There is no check that the columns match the dataset's, which `Dataset.cast()` does perform. The worst case is an **extra** col... | value:https://api.github.com/repos/huggingface/datasets/issues/8453/comments | value:PR_kwDODunzps78iZ8Z | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8453", "html_url": "https://github.com/huggingface/datasets/pull/8453", "diff_url": "https://github.com/huggingface/datasets/pull/8453.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8453.patch", "merged_at": null} | value:2026-08-12T08:10:02Z | value:https://api.github.com/repos/huggingface/datasets/issues/8453/labels{/name} | value:8453 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8452/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/8452 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8452/timeline | json:[] | value:5100498434 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8452 | value:Let a dataset with an Audio column be flattened | value:2026-08-09T02:20:57Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8452/events | value:False | value:`Dataset.flatten()` fails on any dataset that holds a decodable `Audio` column:
```python
from datasets import Dataset, Features, Audio, Value
features = Features({"audio": Audio(), "text": Value("string")})
ds = Dataset.from_dict({"audio": [None], "text": ["a"]}, features=features)
ds.flatten()
```
```
ValueE... | value:https://api.github.com/repos/huggingface/datasets/issues/8452/comments | value:PR_kwDODunzps78iXys | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8452", "html_url": "https://github.com/huggingface/datasets/pull/8452", "diff_url": "https://github.com/huggingface/datasets/pull/8452.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8452.patch", "merged_at": null} | value:2026-08-09T02:20:57Z | value:https://api.github.com/repos/huggingface/datasets/issues/8452/labels{/name} | value:8452 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8451/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/8451 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8451/timeline | json:[] | value:5100493580 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8451 | value:Use the documented default batch_size in the batched IterableDataset exports | value:2026-08-09T02:19:18Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8451/events | value:False | value:`IterableDataset.to_dict()`, `to_pandas()` and `to_polars()` all document `batch_size` as:
> batch_size (`int`, *optional*): The size (number of rows) of the batches if `batched` is `True`. Defaults to `datasets.config.DEFAULT_MAX_BATCH_SIZE`.
but the default is never applied. `batch_size` (`None`) is passed st... | value:https://api.github.com/repos/huggingface/datasets/issues/8451/comments | value:PR_kwDODunzps78iW3q | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8451", "html_url": "https://github.com/huggingface/datasets/pull/8451", "diff_url": "https://github.com/huggingface/datasets/pull/8451.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8451.patch", "merged_at": null} | value:2026-08-12T16:39:27Z | value:https://api.github.com/repos/huggingface/datasets/issues/8451/labels{/name} | value:8451 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8450/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/8450 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8450/timeline | json:[] | value:5100489505 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8450 | value:Raise a clear error for values that cannot be encoded as an image | value:2026-08-09T02:17:47Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8450/events | value:False | value:Putting an unsupported value in an `Image()` column fails with an error from deep inside the encoder rather than a message about the image:
```python
from datasets import Dataset, Features, Image
Dataset.from_dict({"im": [123]}, features=Features({"im": Image()}))
```
```
AttributeError: 'int' object has no at... | value:https://api.github.com/repos/huggingface/datasets/issues/8450/comments | value:PR_kwDODunzps78iWGV | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8450", "html_url": "https://github.com/huggingface/datasets/pull/8450", "diff_url": "https://github.com/huggingface/datasets/pull/8450.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8450.patch", "merged_at": null} | value:2026-08-09T02:17:47Z | value:https://api.github.com/repos/huggingface/datasets/issues/8450/labels{/name} | value:8450 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8449/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/8449 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8449/timeline | json:[] | value:5100489395 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8449 | value:Check the per-split arguments of DatasetDict | value:2026-08-09T02:17:44Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8449/events | value:False | value:`DatasetDict.map`, `filter`, `flatten_indices`, `sort` and `shuffle` take per-split arguments — `cache_file_names`, `indices_cache_file_names`, `seeds`, `generators` — documented as *"You have to provide one ... per dataset in the dataset dictionary"*. None of them checks that, they just index the dict:
```pytho... | value:https://api.github.com/repos/huggingface/datasets/issues/8449/comments | value:PR_kwDODunzps78iWE_ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8449", "html_url": "https://github.com/huggingface/datasets/pull/8449", "diff_url": "https://github.com/huggingface/datasets/pull/8449.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8449.patch", "merged_at": null} | value:2026-08-09T02:17:44Z | value:https://api.github.com/repos/huggingface/datasets/issues/8449/labels{/name} | value:8449 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8448/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/8448 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8448/timeline | json:[] | value:5100481254 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8448 | value:Reject an unknown column in IterableDataset.cast_column | value:2026-08-09T02:15:05Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8448/events | value:False | value:`IterableDataset.cast_column()` writes the feature straight into the features dict:
```python
feature = _fix_for_backward_compatible_features(feature)
info = self._info.copy()
info.features[column] = feature
```
There is no check that `column` is actually a column of the dataset, so a typo doesn't fail — it **i... | value:https://api.github.com/repos/huggingface/datasets/issues/8448/comments | value:PR_kwDODunzps78iUkl | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8448", "html_url": "https://github.com/huggingface/datasets/pull/8448", "diff_url": "https://github.com/huggingface/datasets/pull/8448.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8448.patch", "merged_at": null} | value:2026-08-09T02:15:05Z | value:https://api.github.com/repos/huggingface/datasets/issues/8448/labels{/name} | value:8448 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8447/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/8447 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8447/timeline | json:[] | value:5100479684 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8447 | value:Reject an invalid axis in concatenate_datasets | value:2026-08-09T02:14:35Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8447/events | value:False | value:`concatenate_datasets()` documents `axis` as `{0, 1}`, but does not check it. Any other value silently returns **only the first dataset**:
```python
from datasets import Dataset, concatenate_datasets
a = Dataset.from_dict({"a": [1, 2, 3]})
b = Dataset.from_dict({"b": [4, 5, 6]})
concatenate_datasets([a, b], ax... | value:https://api.github.com/repos/huggingface/datasets/issues/8447/comments | value:PR_kwDODunzps78iUS5 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8447", "html_url": "https://github.com/huggingface/datasets/pull/8447", "diff_url": "https://github.com/huggingface/datasets/pull/8447.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8447.patch", "merged_at": null} | value:2026-08-12T08:00:28Z | value:https://api.github.com/repos/huggingface/datasets/issues/8447/labels{/name} | value:8447 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8446/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/8446 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8446/timeline | json:[] | value:5100471038 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8446 | value:Reject a non-positive batch_size in the Dataset export methods | value:2026-08-09T02:11:35Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8446/events | value:False | value:All the `Dataset.to_*` exports batch with `range(0, len(self), batch_size)`. A negative `batch_size` makes that range empty, so the data is silently dropped — including when writing to a file:
```python
from datasets import Dataset
ds = Dataset.from_dict({"col_1": [0, 1, 2]})
len(list(ds.to_dict(batch_size=-1,... | value:https://api.github.com/repos/huggingface/datasets/issues/8446/comments | value:PR_kwDODunzps78iSqM | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8446", "html_url": "https://github.com/huggingface/datasets/pull/8446", "diff_url": "https://github.com/huggingface/datasets/pull/8446.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8446.patch", "merged_at": null} | value:2026-08-12T07:53:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8446/labels{/name} | value:8446 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8445/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/8445 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8445/timeline | json:[] | value:5100469112 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8445 | value:Fix jax formatting of temporal and decimal columns | value:2026-08-09T02:10:56Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8445/events | value:False | value:Temporal columns are either rejected or **silently corrupted** by the jax format.
### 1. Timestamps, dates, times and decimals raise
```python
import datetime
from datasets import Dataset
ds = Dataset.from_dict({"t": [datetime.datetime(2020, 1, 1)]})
ds.with_format("jax")[0]["t"]
```
```
TypeError: Error inte... | value:https://api.github.com/repos/huggingface/datasets/issues/8445/comments | value:PR_kwDODunzps78iST2 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8445", "html_url": "https://github.com/huggingface/datasets/pull/8445", "diff_url": "https://github.com/huggingface/datasets/pull/8445.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8445.patch", "merged_at": null} | value:2026-08-09T02:10:56Z | value:https://api.github.com/repos/huggingface/datasets/issues/8445/labels{/name} | value:8445 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8444/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/8444 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8444/timeline | json:[] | value:5100467810 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8444 | value:Fix StopIteration when the JSON data files are empty | value:2026-08-09T02:10:30Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8444/events | value:False | value:Feature inference in `json.py` did `next(iter(...))` catching only `FullReadDisallowed`, so an empty `.jsonl` leaked a bare `StopIteration` out of `load_dataset` (the PEP 479 hazard).
This brings `json.py` in line with its neighbours rather than adding a special case: `arrow.py` and `parquet.py` already infer wi... | value:https://api.github.com/repos/huggingface/datasets/issues/8444/comments | value:PR_kwDODunzps78iSDy | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8444", "html_url": "https://github.com/huggingface/datasets/pull/8444", "diff_url": "https://github.com/huggingface/datasets/pull/8444.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8444.patch", "merged_at": null} | value:2026-08-15T13:31:04Z | value:https://api.github.com/repos/huggingface/datasets/issues/8444/labels{/name} | value:8444 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8443/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/8443 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8443/timeline | json:[] | value:5100459074 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8443 | value:Validate by_column and fix the error message in Dataset.batch | value:2026-08-09T02:07:28Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8443/events | value:False | value:Two small problems in `Dataset.batch()`.
**1. The error names the wrong class.**
```python
ds.batch()
# ValueError: IterableDataset.batch() misses `batch_size` or `by_column` argument.
```
`IterableDataset.batch()` has its own copy of the same check, so this occurrence in `arrow_dataset.py` is just a copy-past... | value:https://api.github.com/repos/huggingface/datasets/issues/8443/comments | value:PR_kwDODunzps78iQZs | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8443", "html_url": "https://github.com/huggingface/datasets/pull/8443", "diff_url": "https://github.com/huggingface/datasets/pull/8443.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8443.patch", "merged_at": null} | value:2026-08-12T07:53:47Z | value:https://api.github.com/repos/huggingface/datasets/issues/8443/labels{/name} | value:8443 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8442/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/8442 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8442/timeline | json:[] | value:5100454767 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8442 | value:Accept numpy integers as train_size / test_size in train_test_split | value:2026-08-09T02:06:06Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8442/events | value:False | value:`Dataset.train_test_split()` uses `isinstance(..., int)` / `isinstance(..., float)` to tell "a number of rows" from "a fraction", and rejects everything else:
```python
if train_size is not None and not isinstance(train_size, (int, float)):
raise ValueError(f"Invalid value for train_size: {train_size} of typ... | value:https://api.github.com/repos/huggingface/datasets/issues/8442/comments | value:PR_kwDODunzps78iPlj | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8442", "html_url": "https://github.com/huggingface/datasets/pull/8442", "diff_url": "https://github.com/huggingface/datasets/pull/8442.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8442.patch", "merged_at": null} | value:2026-08-12T11:13:16Z | value:https://api.github.com/repos/huggingface/datasets/issues/8442/labels{/name} | value:8442 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8441/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/8441 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8441/timeline | json:[] | value:5100453439 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8441 | value:Reject a non-positive buffer_size in IterableDataset.shuffle | value:2026-08-09T02:05:37Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8441/events | value:False | value:`IterableDataset.shuffle()` doesn't check `buffer_size`, and the two bad values fail in very different ways.
## A negative buffer_size buffers the whole dataset
`BufferShuffledExamplesIterable.__iter__` fills the buffer with an equality test:
```python
for x in self.ex_iterable:
if len(mem_buffer) == buffe... | value:https://api.github.com/repos/huggingface/datasets/issues/8441/comments | value:PR_kwDODunzps78iPVf | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8441", "html_url": "https://github.com/huggingface/datasets/pull/8441", "diff_url": "https://github.com/huggingface/datasets/pull/8441.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8441.patch", "merged_at": null} | value:2026-08-12T10:32:39Z | value:https://api.github.com/repos/huggingface/datasets/issues/8441/labels{/name} | value:8441 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8440/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/8440 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8440/timeline | json:[] | value:5100449140 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8440 | value:Don't crash when formatting temporal and decimal columns to torch | value:2026-08-09T02:04:04Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8440/events | value:False | value:A dataset with a temporal or decimal column cannot be used with the torch format at all:
```python
import datetime
from datasets import Dataset
ds = Dataset.from_dict({"t": [datetime.datetime(2020, 1, 1)]})
ds.with_format("torch")[0]["t"]
```
```
TypeError: can't convert np.ndarray of type numpy.datetime64.
Th... | value:https://api.github.com/repos/huggingface/datasets/issues/8440/comments | value:PR_kwDODunzps78iOgu | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8440", "html_url": "https://github.com/huggingface/datasets/pull/8440", "diff_url": "https://github.com/huggingface/datasets/pull/8440.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8440.patch", "merged_at": null} | value:2026-08-09T02:06:08Z | value:https://api.github.com/repos/huggingface/datasets/issues/8440/labels{/name} | value:8440 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8439/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/8439 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8439/timeline | json:[] | value:5100448909 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8439 | value:Don't include writer_batch_size in the fingerprint | value:2026-08-09T02:03:59Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8439/events | value:False | value:`writer_batch_size` is documented as
> Number of rows per write operation for the cache file writer. This value is a good trade-off between memory usage during the processing, and processing speed.
It cannot change the data of the resulting dataset — only how it is written. Yet it is part of the fingerprint of ... | value:https://api.github.com/repos/huggingface/datasets/issues/8439/comments | value:PR_kwDODunzps78iOeF | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8439", "html_url": "https://github.com/huggingface/datasets/pull/8439", "diff_url": "https://github.com/huggingface/datasets/pull/8439.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8439.patch", "merged_at": null} | value:2026-08-09T02:03:59Z | value:https://api.github.com/repos/huggingface/datasets/issues/8439/labels{/name} | value:8439 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8438/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/8438 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8438/timeline | json:[] | value:5100439854 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8438 | value:Validate rank and world_size in split_dataset_by_node for IterableDataset | value:2026-08-09T02:00:57Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8438/events | value:False | value:`split_dataset_by_node()` behaves differently on the two dataset types for invalid `rank` / `world_size`. The map-style version delegates to `Dataset.shard()`, which rejects them:
```python
def _split_by_node_map_style_dataset(dataset, rank, world_size):
return dataset.shard(num_shards=world_size, index=rank... | value:https://api.github.com/repos/huggingface/datasets/issues/8438/comments | value:PR_kwDODunzps78iMyJ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8438", "html_url": "https://github.com/huggingface/datasets/pull/8438", "diff_url": "https://github.com/huggingface/datasets/pull/8438.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8438.patch", "merged_at": null} | value:2026-08-12T10:32:41Z | value:https://api.github.com/repos/huggingface/datasets/issues/8438/labels{/name} | value:8438 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8437/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/8437 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8437/timeline | json:[] | value:5100431684 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8437 | value:Raise when a batched filter function returns the wrong number of values | value:2026-08-09T01:58:05Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8437/events | value:False | value:`Dataset.filter(..., batched=True)` builds the kept indices with a `zip`:
```python
indices_array = [i for i, to_keep in zip(indices, mask) if to_keep]
```
`zip` stops at the shorter of its arguments, so a `function` that does not return exactly one value per example is not an error — it silently changes the re... | value:https://api.github.com/repos/huggingface/datasets/issues/8437/comments | value:PR_kwDODunzps78iLQV | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8437", "html_url": "https://github.com/huggingface/datasets/pull/8437", "diff_url": "https://github.com/huggingface/datasets/pull/8437.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8437.patch", "merged_at": null} | value:2026-08-09T01:58:05Z | value:https://api.github.com/repos/huggingface/datasets/issues/8437/labels{/name} | value:8437 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8436/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/8436 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8436/timeline | json:[] | value:5100424782 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8436 | value:Validate seed and generator in Dataset.train_test_split | value:2026-08-09T01:55:34Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8436/events | value:False | value:`Dataset.shuffle()` validates its two randomness parameters:
```python
if seed is not None and generator is not None:
raise ValueError("Both `seed` and `generator` were provided. Please specify just one of them.")
if generator is not None and not isinstance(generator, np.random.Generator):
raise ValueEr... | value:https://api.github.com/repos/huggingface/datasets/issues/8436/comments | value:PR_kwDODunzps78iKAc | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8436", "html_url": "https://github.com/huggingface/datasets/pull/8436", "diff_url": "https://github.com/huggingface/datasets/pull/8436.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8436.patch", "merged_at": null} | value:2026-08-09T01:55:34Z | value:https://api.github.com/repos/huggingface/datasets/issues/8436/labels{/name} | value:8436 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8435/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/8435 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8435/timeline | json:[] | value:5100420317 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8435 | value:Validate num_shards in Dataset.to_iterable_dataset | value:2026-08-09T01:53:55Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8435/events | value:False | value:`Dataset.to_iterable_dataset(num_shards=0)` (or any negative value) returns an `IterableDataset` that silently yields **nothing**:
```python
from datasets import Dataset
ds = Dataset.from_dict({"col_1": [1, 2, 3]})
list(ds.to_iterable_dataset(num_shards=0)) # [] -- no error
list(ds.to_iterable_dataset(num_s... | value:https://api.github.com/repos/huggingface/datasets/issues/8435/comments | value:PR_kwDODunzps78iJJ0 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8435", "html_url": "https://github.com/huggingface/datasets/pull/8435", "diff_url": "https://github.com/huggingface/datasets/pull/8435.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8435.patch", "merged_at": null} | value:2026-08-09T02:22:34Z | value:https://api.github.com/repos/huggingface/datasets/issues/8435/labels{/name} | value:8435 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8434/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/8434 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8434/timeline | json:[] | value:5100419066 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8434 | value:Fix exporting an IterableDataset that has no examples | value:2026-08-09T01:53:27Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8434/events | value:False | value:Every conversion/export method of `IterableDataset` materializes the dataset with `pa.concat_tables()` over its Arrow batches:
```python
table = pa.concat_tables(list(self.with_format("arrow").iter(batch_size=1000)))
return Dataset(table, fingerprint="unset").to_list()
```
If the dataset doesn't yield any examp... | value:https://api.github.com/repos/huggingface/datasets/issues/8434/comments | value:PR_kwDODunzps78iI6a | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8434", "html_url": "https://github.com/huggingface/datasets/pull/8434", "diff_url": "https://github.com/huggingface/datasets/pull/8434.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8434.patch", "merged_at": null} | value:2026-08-09T01:53:27Z | value:https://api.github.com/repos/huggingface/datasets/issues/8434/labels{/name} | value:8434 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8433/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/8433 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8433/timeline | json:[] | value:5100416653 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8433 | value:Validate num_shards in Dataset.save_to_disk | value:2026-08-09T01:52:33Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8433/events | value:False | value:`Dataset.save_to_disk()` accepts `num_shards`, but never checks that it is positive. With `num_shards=0` (or a negative value) it reports success while writing **no data file at all**:
```python
from datasets import Dataset, load_from_disk
ds = Dataset.from_dict({"col_1": [1, 2, 3]})
ds.save_to_disk("dir", num_... | value:https://api.github.com/repos/huggingface/datasets/issues/8433/comments | value:PR_kwDODunzps78iIcD | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8433", "html_url": "https://github.com/huggingface/datasets/pull/8433", "diff_url": "https://github.com/huggingface/datasets/pull/8433.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8433.patch", "merged_at": null} | value:2026-08-09T01:52:33Z | value:https://api.github.com/repos/huggingface/datasets/issues/8433/labels{/name} | value:8433 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8432/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/8432 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8432/timeline | json:[] | value:5100407352 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8432 | value:Fix numpy formatting of a list column whose first row is null | value:2026-08-09T01:49:07Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8432/events | value:False | value:Formatting a list column to numpy or torch crashes when the **first** row is null:
```python
from datasets import Dataset, Features, List, Value
ds = Dataset.from_dict({"c": [None, [3, 4]]}, features=Features({"c": List(Value("int64"))}))
ds.with_format("numpy")[:]["c"]
```
```
AttributeError: 'NoneType' objec... | value:https://api.github.com/repos/huggingface/datasets/issues/8432/comments | value:PR_kwDODunzps78iGpt | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8432", "html_url": "https://github.com/huggingface/datasets/pull/8432", "diff_url": "https://github.com/huggingface/datasets/pull/8432.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8432.patch", "merged_at": null} | value:2026-08-09T01:49:07Z | value:https://api.github.com/repos/huggingface/datasets/issues/8432/labels{/name} | value:8432 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8431/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/8431 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8431/timeline | json:[] | value:5100394795 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8431 | value:Fix the archive member path check against a sibling output directory | value:2026-08-09T01:44:18Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8431/events | value:False | value:The `safemembers()` guards that implement the CVE-2007-4559 mitigation compare the resolved member path to the output directory with a plain `str.startswith()`:
```python
def badpath(path: str, base: str) -> bool:
# joinpath will ignore base if path is absolute
return not resolved(os.path.join(base, path... | value:https://api.github.com/repos/huggingface/datasets/issues/8431/comments | value:PR_kwDODunzps78iEQi | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8431", "html_url": "https://github.com/huggingface/datasets/pull/8431", "diff_url": "https://github.com/huggingface/datasets/pull/8431.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8431.patch", "merged_at": null} | value:2026-08-09T01:44:18Z | value:https://api.github.com/repos/huggingface/datasets/issues/8431/labels{/name} | value:8431 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8430/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/8430 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8430/timeline | json:[] | value:5100363245 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8430 | value:Fix Dataset.flatten with max_depth <= 1 | value:2026-08-09T01:33:04Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8430/events | value:False | value:`Dataset.flatten()` accepts a `max_depth` argument (also exposed through `DatasetDict.flatten()`), but any value `<= 1` crashes:
```python
from datasets import Dataset
Dataset.from_dict({"a": [{"b": {"c": 1}}]}).flatten(max_depth=1)
# UnboundLocalError: cannot access local variable 'depth' where it is not assoc... | value:https://api.github.com/repos/huggingface/datasets/issues/8430/comments | value:PR_kwDODunzps78h-bS | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8430", "html_url": "https://github.com/huggingface/datasets/pull/8430", "diff_url": "https://github.com/huggingface/datasets/pull/8430.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8430.patch", "merged_at": null} | value:2026-08-09T01:48:28Z | value:https://api.github.com/repos/huggingface/datasets/issues/8430/labels{/name} | value:8430 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8429/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/8429 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8429/timeline | json:[] | value:5100349360 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8429 | value:Reject a non-positive batch_size in Dataset.iter | value:2026-08-09T01:28:01Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8429/events | value:False | value:`Dataset.iter(0)` never returns on a dataset without an indices mapping:
```python
from datasets import Dataset
ds = Dataset.from_dict({"i": list(range(5))})
next(ds.iter(0)) # hangs forever
```
The fast path calls `table_iter(self.data, batch_size=0)`, which asks pyarrow for chunks of `max_chunksize=0`; eve... | value:https://api.github.com/repos/huggingface/datasets/issues/8429/comments | value:PR_kwDODunzps78h720 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8429", "html_url": "https://github.com/huggingface/datasets/pull/8429", "diff_url": "https://github.com/huggingface/datasets/pull/8429.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8429.patch", "merged_at": null} | value:2026-08-09T01:28:01Z | value:https://api.github.com/repos/huggingface/datasets/issues/8429/labels{/name} | value:8429 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8428/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/8428 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8428/timeline | json:[] | value:5100348873 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8428 | value:Keep the column order in Features.flatten() | value:2026-08-09T01:27:50Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8428/events | value:False | value:`Features.flatten()` and `Dataset.flatten()` disagree on the resulting column order:
```python
from datasets import Dataset, Features, Value
features = Features({"a": Value("int64"), "s": {"x": Value("string")}, "l": Value("int64")})
ds = Dataset.from_dict({"a": [1], "s": [{"x": "q"}], "l": [2]}, features=featu... | value:https://api.github.com/repos/huggingface/datasets/issues/8428/comments | value:PR_kwDODunzps78h7ww | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8428", "html_url": "https://github.com/huggingface/datasets/pull/8428", "diff_url": "https://github.com/huggingface/datasets/pull/8428.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8428.patch", "merged_at": null} | value:2026-08-09T01:27:50Z | value:https://api.github.com/repos/huggingface/datasets/issues/8428/labels{/name} | value:8428 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8427/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/8427 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8427/timeline | json:[] | value:5100334562 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8427 | value:Validate the arguments of IterableDataset.shard | value:2026-08-09T01:23:05Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8427/events | value:False | value:`IterableDataset.shard()` doesn't validate its arguments, so invalid calls either return the wrong data silently or crash with a low-level error from an internal helper. `Dataset.shard()` already validates the index.
```python
ds = Dataset.from_dict({"a": range(20)}).to_iterable_dataset(num_shards=4)
ds.shard(n... | value:https://api.github.com/repos/huggingface/datasets/issues/8427/comments | value:PR_kwDODunzps78h5C5 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8427", "html_url": "https://github.com/huggingface/datasets/pull/8427", "diff_url": "https://github.com/huggingface/datasets/pull/8427.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8427.patch", "merged_at": null} | value:2026-08-09T01:23:05Z | value:https://api.github.com/repos/huggingface/datasets/issues/8427/labels{/name} | value:8427 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8426/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/8426 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8426/timeline | json:[] | value:5100333579 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8426 | value:Fix casting a chunked array that has no chunk | value:2026-08-09T01:22:43Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8426/events | value:False | value:A batched `map` on an arrow-formatted dataset crashes when the function filters out every row of a batch and the output features are given explicitly:
```python
import pyarrow.compute as pc
from datasets import Dataset, Features, Value
ds = Dataset.from_dict({"a": [1, 2, 3, 4]}).with_format("arrow")
ds.map(
... | value:https://api.github.com/repos/huggingface/datasets/issues/8426/comments | value:PR_kwDODunzps78h42r | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8426", "html_url": "https://github.com/huggingface/datasets/pull/8426", "diff_url": "https://github.com/huggingface/datasets/pull/8426.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8426.patch", "merged_at": null} | value:2026-08-09T01:22:43Z | value:https://api.github.com/repos/huggingface/datasets/issues/8426/labels{/name} | value:8426 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8425/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/8425 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8425/timeline | json:[] | value:5100317104 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8425 | value:Return a new Dataset when transforming a dataset with no rows | value:2026-08-09T01:16:45Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8425/events | value:False | value:`Dataset.map`, `.filter`, `.select`, `.sort`, `.shuffle` and `.train_test_split` short-circuit when the dataset has no rows and return `self` — the very same object. `train_test_split` even returns that same object as *both* splits.
Because `set_format` / `set_transform` / `reset_format` / `add_faiss_index` muta... | value:https://api.github.com/repos/huggingface/datasets/issues/8425/comments | value:PR_kwDODunzps78h1xs | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8425", "html_url": "https://github.com/huggingface/datasets/pull/8425", "diff_url": "https://github.com/huggingface/datasets/pull/8425.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8425.patch", "merged_at": null} | value:2026-08-09T01:16:45Z | value:https://api.github.com/repos/huggingface/datasets/issues/8425/labels{/name} | value:8425 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8424/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/8424 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8424/timeline | json:[] | value:5100302571 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8424 | value:Encode the column in Dataset.add_column when a feature is given | value:2026-08-09T01:11:48Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8424/events | value:False | value:`Dataset.add_column()` takes a `feature` argument documented as the "column datatype", but it is passed directly as the arrow schema of `InMemoryTable.from_pydict()`. That only works when the raw Python values already match the feature's *storage* type, so any feature that encodes its values fails:
```python
fro... | value:https://api.github.com/repos/huggingface/datasets/issues/8424/comments | value:PR_kwDODunzps78hzF4 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8424", "html_url": "https://github.com/huggingface/datasets/pull/8424", "diff_url": "https://github.com/huggingface/datasets/pull/8424.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8424.patch", "merged_at": null} | value:2026-08-09T01:11:48Z | value:https://api.github.com/repos/huggingface/datasets/issues/8424/labels{/name} | value:8424 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8423/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/8423 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8423/timeline | json:[] | value:5100294829 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8423 | value:Fix keep_in_memory=True in Dataset.sort and Dataset.train_test_split | value:2026-08-09T01:09:05Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8423/events | value:False | value:`Dataset.sort()` and `Dataset.train_test_split()` raise as soon as `keep_in_memory=True` is used on a dataset that has cache files — i.e. on anything returned by `load_dataset()` or `load_from_disk()`:
```python
from datasets import Dataset
Dataset.from_dict({"col_1": [3, 1, 2]}).save_to_disk("ds")
ds = Dataset... | value:https://api.github.com/repos/huggingface/datasets/issues/8423/comments | value:PR_kwDODunzps78hxqF | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8423", "html_url": "https://github.com/huggingface/datasets/pull/8423", "diff_url": "https://github.com/huggingface/datasets/pull/8423.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8423.patch", "merged_at": null} | value:2026-08-09T01:09:05Z | value:https://api.github.com/repos/huggingface/datasets/issues/8423/labels{/name} | value:8423 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8422/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/8422 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8422/timeline | json:[] | value:5100293775 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8422 | value:Fix to_pandas() on a dataset with no rows and an ArrayXD column | value:2026-08-09T01:08:42Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8422/events | value:False | value:Converting a dataset that has no rows and an `ArrayXD` column to pandas raises:
```python
import datasets
from datasets import Dataset, Features, Array2D
features = Features({"foo": Array2D(shape=(2, 2), dtype="float32")})
ds = Dataset.from_dict({"foo": []}, features=features)
ds.to_pandas()
```
```
pyarrow.li... | value:https://api.github.com/repos/huggingface/datasets/issues/8422/comments | value:PR_kwDODunzps78hxcW | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8422", "html_url": "https://github.com/huggingface/datasets/pull/8422", "diff_url": "https://github.com/huggingface/datasets/pull/8422.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8422.patch", "merged_at": null} | value:2026-08-09T01:08:42Z | value:https://api.github.com/repos/huggingface/datasets/issues/8422/labels{/name} | value:8422 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8421/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/8421 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8421/timeline | json:[] | value:5100293456 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8421 | value:Validate the columns in IterableDataset.rename_columns | value:2026-08-09T01:08:36Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8421/events | value:False | value:`IterableDataset.rename_columns()` doesn't check the column names it is given, so when `info.features` is known it can return a dataset whose `features` contradict the examples it yields. The error is only raised much later, by `_rename_columns_fn()`, when the dataset is iterated.
```python
ds = Dataset.from_dic... | value:https://api.github.com/repos/huggingface/datasets/issues/8421/comments | value:PR_kwDODunzps78hxYV | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8421", "html_url": "https://github.com/huggingface/datasets/pull/8421", "diff_url": "https://github.com/huggingface/datasets/pull/8421.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8421.patch", "merged_at": null} | value:2026-08-09T01:08:36Z | value:https://api.github.com/repos/huggingface/datasets/issues/8421/labels{/name} | value:8421 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8420/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/8420 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8420/timeline | json:[] | value:5100282369 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8420 | value:Keep the dtype of an empty column when formatting to numpy or torch | value:2026-08-09T01:05:16Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8420/events | value:False | value:An empty column silently changes dtype when formatted to numpy or torch:
```python
from datasets import Dataset
ds = Dataset.from_dict({"a": [1, 2, 3]}).with_format("torch")
print(ds[:]["a"].dtype) # torch.int64
print(ds.filter(lambda x: False)[:]["a"].dtype) # torch.float32 <-- expec... | value:https://api.github.com/repos/huggingface/datasets/issues/8420/comments | value:PR_kwDODunzps78hvMN | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8420", "html_url": "https://github.com/huggingface/datasets/pull/8420", "diff_url": "https://github.com/huggingface/datasets/pull/8420.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8420.patch", "merged_at": null} | value:2026-08-09T01:05:16Z | value:https://api.github.com/repos/huggingface/datasets/issues/8420/labels{/name} | value:8420 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8419/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/8419 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8419/timeline | json:[] | value:5100271993 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8419 | value:Fix label inference for folder-based builders on archives | value:2026-08-09T01:01:47Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8419/events | value:False | value:`imagefolder` / `audiofolder` / `videofolder` datasets whose media files live inside a ZIP archive get the wrong labels, and usually fail to build at all.
## The bug
In `_split_generators()` the entries of `files` have two different shapes:
```python
# loose files: (original_file, None, downloaded_file)
files... | value:https://api.github.com/repos/huggingface/datasets/issues/8419/comments | value:PR_kwDODunzps78htLy | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8419", "html_url": "https://github.com/huggingface/datasets/pull/8419", "diff_url": "https://github.com/huggingface/datasets/pull/8419.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8419.patch", "merged_at": null} | value:2026-08-09T01:01:47Z | value:https://api.github.com/repos/huggingface/datasets/issues/8419/labels{/name} | value:8419 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8418/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/8418 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8418/timeline | json:[] | value:5100259233 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8418 | value:Fix DataFrame masking on ArrayXD columns with an integer value type | value:2026-08-09T00:57:20Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8418/events | value:False | value:Selecting rows with a boolean mask on a `DataFrame` that has an `Array2D`/`Array3D`/... column with an integer value type raises:
```python
import datasets, numpy as np
features = datasets.Features({"id": datasets.Value("int64"), "foo": datasets.Array2D(dtype="int32", shape=(2, 2))})
ds = datasets.Dataset.from_... | value:https://api.github.com/repos/huggingface/datasets/issues/8418/comments | value:PR_kwDODunzps78hq0h | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8418", "html_url": "https://github.com/huggingface/datasets/pull/8418", "diff_url": "https://github.com/huggingface/datasets/pull/8418.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8418.patch", "merged_at": null} | value:2026-08-09T00:57:20Z | value:https://api.github.com/repos/huggingface/datasets/issues/8418/labels{/name} | value:8418 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8417/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/8417 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8417/timeline | json:[] | value:5098273405 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8417 | value:Shutdown ThreadPoolExecutor in CyclingMultiSourcesExamplesIterable | value:2026-08-08T15:58:26Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8417/events | value:False | value:Add missing executor.shutdown(wait=True) call in the finally block of _iter_arrow(), matching the behavior in __iter__ to ensure threads are shut down before cleanup and sleep. | value:https://api.github.com/repos/huggingface/datasets/issues/8417/comments | value:PR_kwDODunzps78brrn | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8417", "html_url": "https://github.com/huggingface/datasets/pull/8417", "diff_url": "https://github.com/huggingface/datasets/pull/8417.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8417.patch", "merged_at": null} | value:2026-08-08T15:58:26Z | value:https://api.github.com/repos/huggingface/datasets/issues/8417/labels{/name} | value:8417 | json:{"login": "HassanRady", "id": 48225007, "node_id": "MDQ6VXNlcjQ4MjI1MDA3", "avatar_url": "https://avatars.githubusercontent.com/u/48225007?v=4", "gravatar_id": "", "url": "https://api.github.com/users/HassanRady", "html_url": "https://github.com/HassanRady", "followers_url": "https://api.github.com/users/HassanRad... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8416/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/8416 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8416/timeline | json:[] | value:5097697849 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8416 | value:Fix cryptic TypeError from ClassLabel.int2str on a string input | value:2026-08-08T13:27:22Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8416/events | value:False | value:`ClassLabel.int2str("1")` raises a cryptic `TypeError: '<=' not supported between instances of 'int' and 'str'` instead of a clear error.
`int2str` guards its input with `isinstance(values, Iterable)`, but a `str` is iterable, so a string is not wrapped as a single value; the code then iterates over the string's... | value:https://api.github.com/repos/huggingface/datasets/issues/8416/comments | value:PR_kwDODunzps78Z-OF | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8416", "html_url": "https://github.com/huggingface/datasets/pull/8416", "diff_url": "https://github.com/huggingface/datasets/pull/8416.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8416.patch", "merged_at": null} | value:2026-08-15T11:16:29Z | value:https://api.github.com/repos/huggingface/datasets/issues/8416/labels{/name} | value:8416 | json:{"login": "hdimer", "id": 94395445, "node_id": "U_kgDOBaBcNQ", "avatar_url": "https://avatars.githubusercontent.com/u/94395445?v=4", "gravatar_id": "", "url": "https://api.github.com/users/hdimer", "html_url": "https://github.com/hdimer", "followers_url": "https://api.github.com/users/hdimer/followers", "following... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8415/reactions", "total_count": 1, "+1": 1, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 0, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8415 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8415/timeline | json:[] | value:5091230885 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8415 | value:Add initial devcontainer configuration | value:2026-08-07T14:44:30Z | value:NONE | json:{"login": "mo74m3ed", "id": 71875716, "node_id": "MDQ6VXNlcjcxODc1NzE2", "avatar_url": "https://avatars.githubusercontent.com/u/71875716?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mo74m3ed", "html_url": "https://github.com/mo74m3ed", "followers_url": "https://api.github.com/users/mo74m3ed/follow... | value:https://api.github.com/repos/huggingface/datasets/issues/8415/events | value:False | value:2026-08-21T18:15:16Z | value:https://api.github.com/repos/huggingface/datasets/issues/8415/comments | value:PR_kwDODunzps78Fm6_ | value:closed | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8415", "html_url": "https://github.com/huggingface/datasets/pull/8415", "diff_url": "https://github.com/huggingface/datasets/pull/8415.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8415.patch", "merged_at": null} | value:2026-08-21T18:15:16Z | value:https://api.github.com/repos/huggingface/datasets/issues/8415/labels{/name} | value:8415 | json:{"login": "mo74m3ed", "id": 71875716, "node_id": "MDQ6VXNlcjcxODc1NzE2", "avatar_url": "https://avatars.githubusercontent.com/u/71875716?v=4", "gravatar_id": "", "url": "https://api.github.com/users/mo74m3ed", "html_url": "https://github.com/mo74m3ed", "followers_url": "https://api.github.com/users/mo74m3ed/follow... | value:False | true | ||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8414/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/8414 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8414/timeline | json:[] | value:5090218572 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8414 | value:Fix offline simulation with huggingface_hub 1.27 | value:2026-08-07T12:39:32Z | 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/8414/events | value:False | value:2026-08-10T19:31:39Z | value:## Summary
Surfaced by the `huggingface_hub` v1.27.0.rc0 RC test branch, where 6 tests in `tests/test_load.py` fail: [ci-test-huggingface-hub-1.27.0.rc0-release](https://github.com/huggingface/datasets/compare/main...ci-test-huggingface-hub-1.27.0.rc0-release).
They're all offline-simulation tests, and the fail... | value:https://api.github.com/repos/huggingface/datasets/issues/8414/comments | value:PR_kwDODunzps78CWen | value:closed | value:2 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8414", "html_url": "https://github.com/huggingface/datasets/pull/8414", "diff_url": "https://github.com/huggingface/datasets/pull/8414.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8414.patch", "merged_at": "2026-08-10T19:31:39Z"} | value:2026-08-10T19:31:41Z | value:https://api.github.com/repos/huggingface/datasets/issues/8414/labels{/name} | value:8414 | json:{"login": "hanouticelina", "id": 36770234, "node_id": "MDQ6VXNlcjM2NzcwMjM0", "avatar_url": "https://avatars.githubusercontent.com/u/36770234?v=4", "gravatar_id": "", "url": "https://api.github.com/users/hanouticelina", "html_url": "https://github.com/hanouticelina", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8412/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/8412 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8412/timeline | json:[] | value:5084411761 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8412 | value:Fix IndexError on negative list indices for datasets with an indices mapping | value:2026-08-06T20:02:13Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8412/events | value:False | value:## Summary
In `src/datasets/formatting/formatting.py`, `_query_table_with_indices_mapping`'s iterable branch passes each index directly to `indices.fast_slice(i, 1)`:
```python
return _query_table(table, [indices.fast_slice(i, 1).column(0)[0].as_py() for i in key])
```
`fast_slice` raises `IndexError("Offset m... | value:https://api.github.com/repos/huggingface/datasets/issues/8412/comments | value:PR_kwDODunzps77v1RT | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8412", "html_url": "https://github.com/huggingface/datasets/pull/8412", "diff_url": "https://github.com/huggingface/datasets/pull/8412.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8412.patch", "merged_at": null} | value:2026-08-06T20:02:13Z | value:https://api.github.com/repos/huggingface/datasets/issues/8412/labels{/name} | value:8412 | json:{"login": "uttam12331", "id": 159913047, "node_id": "U_kgDOCYgUVw", "avatar_url": "https://avatars.githubusercontent.com/u/159913047?v=4", "gravatar_id": "", "url": "https://api.github.com/users/uttam12331", "html_url": "https://github.com/uttam12331", "followers_url": "https://api.github.com/users/uttam12331/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8411/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/8411 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8411/timeline | json:[] | value:5083951373 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8411 | value:Restore the _legacy_no_dict_keys_sorting branch in Pickler._batch_setitems | value:2026-08-06T18:40:51Z | value:CONTRIBUTOR | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | value:https://api.github.com/repos/huggingface/datasets/issues/8411/events | value:False | value:2026-09-06T00:05:48Z | value:### Root cause
`Pickler._legacy_no_dict_keys_sorting` lost its only reader in #7817. That PR rewrote `_batch_setitems` to accept the argument Python 3.14 passes, and in the same hunk dropped the early return:
```diff
- def _batch_setitems(self, items):
- if self._legacy_no_dict_keys_sorting:
- ... | value:https://api.github.com/repos/huggingface/datasets/issues/8411/comments | value:PR_kwDODunzps77uYL- | value:closed | value:2 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8411", "html_url": "https://github.com/huggingface/datasets/pull/8411", "diff_url": "https://github.com/huggingface/datasets/pull/8411.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8411.patch", "merged_at": null} | value:2026-09-06T00:05:48Z | value:https://api.github.com/repos/huggingface/datasets/issues/8411/labels{/name} | value:8411 | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8410/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/8410 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8410/timeline | json:[] | value:5083948009 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8410 | value:_legacy_no_dict_keys_sorting has no reader since #7817, so the 2.15.0 legacy cache lookup computes the wrong config_id | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-06T18:40:14Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8410/events | value:False | value:### Describe the bug
`Pickler._legacy_no_dict_keys_sorting` has a definition and a writer, but since #7817 it has no reader, so the flag does nothing.
`DatasetBuilder._check_legacy_cache2` patches it on to recompute the `config_id` that 2.15.0 would have produced, because 2.15.0 hashed dicts without sorting the... | value:https://api.github.com/repos/huggingface/datasets/issues/8410/comments | value:I_kwDODunzps8AAAABLwbj6Q | value:open | value:0 | value:2026-08-06T18:40:14Z | value:https://api.github.com/repos/huggingface/datasets/issues/8410/labels{/name} | value:8410 | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8409/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/8409 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8409/timeline | json:[] | value:5083087639 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8409 | value:docs: fix broken audio link in mapstyle vs iterable guide | value:2026-08-06T16:13:40Z | value:NONE | json:{"login": "latent-9", "id": 296084221, "node_id": "U_kgDOEaXi_Q", "avatar_url": "https://avatars.githubusercontent.com/u/296084221?v=4", "gravatar_id": "", "url": "https://api.github.com/users/latent-9", "html_url": "https://github.com/latent-9", "followers_url": "https://api.github.com/users/latent-9/followers", ... | value:https://api.github.com/repos/huggingface/datasets/issues/8409/events | value:False | value:2026-08-15T08:34:39Z | value:The "map-style vs iterable" guide links the four loading guides on one line, but the `audio` link has a stray `]` inside its target (`./audio_load]`), which makes the Markdown link malformed and breaks it.
The three sibling links on the same line are correct (`./tabular_load`, `./nlp_load`, `./image_load`), and ... | value:https://api.github.com/repos/huggingface/datasets/issues/8409/comments | value:PR_kwDODunzps77roh- | value:closed | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8409", "html_url": "https://github.com/huggingface/datasets/pull/8409", "diff_url": "https://github.com/huggingface/datasets/pull/8409.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8409.patch", "merged_at": null} | value:2026-08-15T08:34:39Z | value:https://api.github.com/repos/huggingface/datasets/issues/8409/labels{/name} | value:8409 | json:{"login": "latent-9", "id": 296084221, "node_id": "U_kgDOEaXi_Q", "avatar_url": "https://avatars.githubusercontent.com/u/296084221?v=4", "gravatar_id": "", "url": "https://api.github.com/users/latent-9", "html_url": "https://github.com/latent-9", "followers_url": "https://api.github.com/users/latent-9/followers", ... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8408/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/8408 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8408/timeline | json:[] | value:5080309439 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8408 | value:Save the shuffle buffer content in the state dict so resuming is exact | value:2026-08-06T10:41:20Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8408/events | value:False | value:## Overview
There are two problems when resuming a shuffled iterable dataset. The known one from the warning and the discussion in #7901 is that the shuffle buffer content isn't saved in the state dict so you lose roughly `buffer_size` examples every resume.
The second is you can get an example twice. The `... | value:https://api.github.com/repos/huggingface/datasets/issues/8408/comments | value:PR_kwDODunzps77ilJy | value:open | value:3 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8408", "html_url": "https://github.com/huggingface/datasets/pull/8408", "diff_url": "https://github.com/huggingface/datasets/pull/8408.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8408.patch", "merged_at": null} | value:2026-10-23T16:47:41Z | value:https://api.github.com/repos/huggingface/datasets/issues/8408/labels{/name} | value:8408 | json:{"login": "thejamesgore", "id": 83005220, "node_id": "MDQ6VXNlcjgzMDA1MjIw", "avatar_url": "https://avatars.githubusercontent.com/u/83005220?v=4", "gravatar_id": "", "url": "https://api.github.com/users/thejamesgore", "html_url": "https://github.com/thejamesgore", "followers_url": "https://api.github.com/users/the... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8407/reactions", "total_count": 1, "+1": 0, "-1": 0, "laugh": 0, "hooray": 0, "confused": 0, "heart": 0, "rocket": 1, "eyes": 0} | json:[] | value:https://github.com/huggingface/datasets/pull/8407 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8407/timeline | json:[] | value:5077929707 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8407 | value:Optimize PyArrow block shuffling with dynamic rebatched chunk interleaving | value:2026-08-06T05:24:23Z | 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/8407/events | value:False | value:2026-09-07T14:03:29Z | value:When streaming PyArrow datasets with shuffling enabled (`buffer_size > 0`), the `BufferShuffledExamplesIterable._iter_arrow()` method historically relied on a Python-level `list` to maintain the shuffle buffer, replacing elements one-by-one. In a multi-worker `DataLoader` environment, this per-row generator appro... | value:https://api.github.com/repos/huggingface/datasets/issues/8407/comments | value:PR_kwDODunzps77a8Eg | value:closed | value:6 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8407", "html_url": "https://github.com/huggingface/datasets/pull/8407", "diff_url": "https://github.com/huggingface/datasets/pull/8407.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8407.patch", "merged_at": "2026-09-07T14:03:29Z"} | value:2026-10-07T07:54:31Z | value:https://api.github.com/repos/huggingface/datasets/issues/8407/labels{/name} | value:8407 | json:{"login": "yuxin00j", "id": 255896926, "node_id": "U_kgDOD0CtXg", "avatar_url": "https://avatars.githubusercontent.com/u/255896926?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yuxin00j", "html_url": "https://github.com/yuxin00j", "followers_url": "https://api.github.com/users/yuxin00j/followers", ... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8406/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/8406 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8406/timeline | json:[] | value:5076886721 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8406 | value:Preserve nanosecond precision in Python dataset construction | value:2026-08-06T02:09:38Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8406/events | value:False | value:## What it is
Preserves exact nanosecond precision for pandas timestamps and timedeltas when users declare `Value("timestamp[ns]")` or `Value("duration[ns]")`.
Previously, `Dataset.from_dict`, `Dataset.from_list`, and `Dataset.map` reported a nanosecond feature while silently storing values truncated to microse... | value:https://api.github.com/repos/huggingface/datasets/issues/8406/comments | value:PR_kwDODunzps77XnBU | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8406", "html_url": "https://github.com/huggingface/datasets/pull/8406", "diff_url": "https://github.com/huggingface/datasets/pull/8406.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8406.patch", "merged_at": null} | value:2026-08-06T02:43:17Z | value:https://api.github.com/repos/huggingface/datasets/issues/8406/labels{/name} | value:8406 | json:{"login": "sophianggan", "id": 237162263, "node_id": "U_kgDODiLPFw", "avatar_url": "https://avatars.githubusercontent.com/u/237162263?v=4", "gravatar_id": "", "url": "https://api.github.com/users/sophianggan", "html_url": "https://github.com/sophianggan", "followers_url": "https://api.github.com/users/sophianggan/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8405/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/8405 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8405/timeline | json:[] | value:5074026078 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8405 | value:Match Dataset.map's on_mixed_types default in DatasetDict.map | value:2026-08-05T18:31:26Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8405/events | value:False | value:### Problem
`Dataset.map` defaults `on_mixed_types` to `"use_json"`, so a mapped column with no
predefined schema becomes a `Json()` feature instead of failing type inference.
`DatasetDict.map` defaults the same parameter to `None` and forwards it to every split
unconditionally, so wrapping the splits in a `Data... | value:https://api.github.com/repos/huggingface/datasets/issues/8405/comments | value:PR_kwDODunzps77OaCx | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8405", "html_url": "https://github.com/huggingface/datasets/pull/8405", "diff_url": "https://github.com/huggingface/datasets/pull/8405.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8405.patch", "merged_at": null} | value:2026-10-21T15:15:10Z | value:https://api.github.com/repos/huggingface/datasets/issues/8405/labels{/name} | value:8405 | json:{"login": "vineethsaivs", "id": 42468190, "node_id": "MDQ6VXNlcjQyNDY4MTkw", "avatar_url": "https://avatars.githubusercontent.com/u/42468190?v=4", "gravatar_id": "", "url": "https://api.github.com/users/vineethsaivs", "html_url": "https://github.com/vineethsaivs", "followers_url": "https://api.github.com/users/vin... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8404/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/8404 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8404/timeline | json:[] | value:5070831360 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8404 | value:Fix broken links in documentation | value:2026-08-05T12:27:50Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8404/events | value:False | value:Several documentation pages link to URLs that now return 404. This PR updates them to their current locations.
| File | Old (404) | New |
|------|-----------|-----|
| `object_detection.mdx` | `albumentations.ai/docs/examples/example_bboxes/` | `albumentations.ai/docs/examples/example-bboxes/` |
| `image_dataset.... | value:https://api.github.com/repos/huggingface/datasets/issues/8404/comments | value:PR_kwDODunzps77EBK8 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8404", "html_url": "https://github.com/huggingface/datasets/pull/8404", "diff_url": "https://github.com/huggingface/datasets/pull/8404.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8404.patch", "merged_at": null} | value:2026-08-05T12:27:50Z | value:https://api.github.com/repos/huggingface/datasets/issues/8404/labels{/name} | value:8404 | json:{"login": "richboyneedcash", "id": 273099414, "node_id": "U_kgDOEEcqlg", "avatar_url": "https://avatars.githubusercontent.com/u/273099414?v=4", "gravatar_id": "", "url": "https://api.github.com/users/richboyneedcash", "html_url": "https://github.com/richboyneedcash", "followers_url": "https://api.github.com/users/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8403/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/8403 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8403/timeline | json:[] | value:5064557555 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8403 | value:Fix pandas formatting when a decodable column is excluded by columns= | value:2026-08-04T20:21:35Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8403/events | value:False | value:The pandas decoder builds its transform map from all of `self.features`, but the frame has already been projected to the format's `columns`:
```python
ds.with_format("pandas", columns=["label"])[0]
# KeyError: "Label(s) ['image'] do not exist"
```
Scoped to pandas — `None`, `numpy`, `torch` and `arrow` all hand... | value:https://api.github.com/repos/huggingface/datasets/issues/8403/comments | value:PR_kwDODunzps76v2d3 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8403", "html_url": "https://github.com/huggingface/datasets/pull/8403", "diff_url": "https://github.com/huggingface/datasets/pull/8403.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8403.patch", "merged_at": null} | value:2026-08-04T20:21:35Z | value:https://api.github.com/repos/huggingface/datasets/issues/8403/labels{/name} | value:8403 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8402/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/8402 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8402/timeline | json:[] | value:5064555950 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8402 | value:Fix PolarsFeaturesDecoder.decode_row | value:2026-08-04T20:21:21Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8402/events | value:False | value:`PolarsFeaturesDecoder.decode_row` is a verbatim copy of the pandas implementation, but polars' `map_rows` takes a callable rather than a dict of column transforms:
```python
ds.with_format("polars")[0]
# TypeError: 'dict' object is not callable
```
So `with_format("polars")` is unusable on any dataset with an ... | value:https://api.github.com/repos/huggingface/datasets/issues/8402/comments | value:PR_kwDODunzps76v2Jl | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8402", "html_url": "https://github.com/huggingface/datasets/pull/8402", "diff_url": "https://github.com/huggingface/datasets/pull/8402.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8402.patch", "merged_at": null} | value:2026-08-04T20:21:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8402/labels{/name} | value:8402 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8401/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/8401 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8401/timeline | json:[] | value:5064554167 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8401 | value:Allow reading a split with no examples | value:2026-08-04T20:21:07Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8401/events | value:False | value:`make_file_instructions` skips every file whose take is 0, so `BaseReader.read` receives an empty file list and raises — which means one empty split makes the entire `DatasetDict` unloadable:
```python
load_dataset("json", data_files={"train": [...], "test": []})
# ValueError: Instruction "test" corresponds to n... | value:https://api.github.com/repos/huggingface/datasets/issues/8401/comments | value:PR_kwDODunzps76v1yt | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8401", "html_url": "https://github.com/huggingface/datasets/pull/8401", "diff_url": "https://github.com/huggingface/datasets/pull/8401.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8401.patch", "merged_at": null} | value:2026-08-04T20:21:07Z | value:https://api.github.com/repos/huggingface/datasets/issues/8401/labels{/name} | value:8401 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8400/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/8400 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8400/timeline | json:[] | value:5064552446 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8400 | value:Fix align_labels_with_mapping on a list of ClassLabel | value:2026-08-04T20:20:52Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8400/events | value:False | value:#7634 dropped `Sequence` from the `.features` import but left two references in `align_labels_with_mapping`, so every non-plain-`ClassLabel` column raises:
```python
ds.align_labels_with_mapping({"O": 0, "LOC": 1, "PER": 2}, "ner_tags")
# NameError: name 'Sequence' is not defined
```
That covers the token-class... | value:https://api.github.com/repos/huggingface/datasets/issues/8400/comments | value:PR_kwDODunzps76v1cs | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8400", "html_url": "https://github.com/huggingface/datasets/pull/8400", "diff_url": "https://github.com/huggingface/datasets/pull/8400.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8400.patch", "merged_at": null} | value:2026-08-09T01:49:11Z | value:https://api.github.com/repos/huggingface/datasets/issues/8400/labels{/name} | value:8400 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8399/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/8399 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8399/timeline | json:[] | value:5064550459 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8399 | value:Fix interleave_datasets with all_exhausted when a dataset is empty | value:2026-08-04T20:20:35Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8399/events | value:False | value:With `stopping_strategy="all_exhausted"` and no `probabilities`, the indices come from `np.arange(max(lengths)) % lengths`. A zero-length input makes that a modulo by zero, which yields index 0; after the source offset is added it is either out of range or a valid index into a *different* dataset:
```python
inte... | value:https://api.github.com/repos/huggingface/datasets/issues/8399/comments | value:PR_kwDODunzps76v1Cv | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8399", "html_url": "https://github.com/huggingface/datasets/pull/8399", "diff_url": "https://github.com/huggingface/datasets/pull/8399.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8399.patch", "merged_at": null} | value:2026-08-04T20:20:35Z | value:https://api.github.com/repos/huggingface/datasets/issues/8399/labels{/name} | value:8399 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8398/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/8398 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8398/timeline | json:[] | value:5064548613 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8398 | value:Fix casting a sliced list array to a fixed size list | value:2026-08-04T20:20:20Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8398/events | value:False | value:`array_cast` and `cast_array_to_feature` build the child window for a fixed-size list as `array.values[array.offset * length : ...]`. But `.values` is the **unsliced** child array, so that arithmetic is only correct when every earlier row holds exactly `length` items. `Dataset.cast` slices into batches, so this f... | value:https://api.github.com/repos/huggingface/datasets/issues/8398/comments | value:PR_kwDODunzps76v0qb | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8398", "html_url": "https://github.com/huggingface/datasets/pull/8398", "diff_url": "https://github.com/huggingface/datasets/pull/8398.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8398.patch", "merged_at": null} | value:2026-08-04T20:20:20Z | value:https://api.github.com/repos/huggingface/datasets/issues/8398/labels{/name} | value:8398 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8397/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/8397 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8397/timeline | json:[] | value:5064546506 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8397 | value:Support empty input shards in ArrowBasedBuilder | value:2026-08-04T20:20:03Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8397/events | value:False | value:`ArrowBasedBuilder._prepare_split_single` assumes every input shard yields at least one table, so `original_shard_lengths` is indexed out of range when a shard yields nothing:
```python
load_dataset("json", data_files=["00.jsonl", "01.jsonl"]) # 00.jsonl is empty
# DatasetGenerationError
# cause: IndexError:... | value:https://api.github.com/repos/huggingface/datasets/issues/8397/comments | value:PR_kwDODunzps76v0P0 | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8397", "html_url": "https://github.com/huggingface/datasets/pull/8397", "diff_url": "https://github.com/huggingface/datasets/pull/8397.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8397.patch", "merged_at": null} | value:2026-08-04T20:20:03Z | value:https://api.github.com/repos/huggingface/datasets/issues/8397/labels{/name} | value:8397 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8396/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/8396 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8396/timeline | json:[] | value:5063822443 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8396 | value:Fix malformed audio link in mapstyle vs iterable docs | value:2026-08-04T18:48:43Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8396/events | value:False | value:# What does this PR do?
Fixes a malformed link in `docs/source/about_mapstyle_vs_iterable.mdx`.
The audio guide link was written as `[audio](./audio_load])` (a stray closing parenthesis inside the link target), so it resolved to a non-existent path. This PR removes the stray `)` so the link points to the actual... | value:https://api.github.com/repos/huggingface/datasets/issues/8396/comments | value:PR_kwDODunzps76teEZ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8396", "html_url": "https://github.com/huggingface/datasets/pull/8396", "diff_url": "https://github.com/huggingface/datasets/pull/8396.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8396.patch", "merged_at": null} | value:2026-08-04T18:48:43Z | value:https://api.github.com/repos/huggingface/datasets/issues/8396/labels{/name} | value:8396 | json:{"login": "seedaylight", "id": 244760094, "node_id": "U_kgDODpa-Hg", "avatar_url": "https://avatars.githubusercontent.com/u/244760094?v=4", "gravatar_id": "", "url": "https://api.github.com/users/seedaylight", "html_url": "https://github.com/seedaylight", "followers_url": "https://api.github.com/users/seedaylight/... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8395/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/8395 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8395/timeline | json:[] | value:5060759129 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8395 | value:Forward iter_arrow, is_typed and features from RepeatExamplesIterable | value:2026-08-04T13:01:32Z | 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/8395/events | value:False | value:2026-08-31T09:38:01Z | value:`RepeatExamplesIterable` never forwarded `iter_arrow`, `is_typed` or `features` to the iterable it wraps, so `.repeat()` stranded an Arrow-capable child behind a non-Arrow parent: iteration fell back to the per-example path and the dtypes were re-inferred from Python objects.
This copies the three-property block... | value:https://api.github.com/repos/huggingface/datasets/issues/8395/comments | value:PR_kwDODunzps76jdoX | value:closed | value:3 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8395", "html_url": "https://github.com/huggingface/datasets/pull/8395", "diff_url": "https://github.com/huggingface/datasets/pull/8395.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8395.patch", "merged_at": "2026-08-31T09:38:01Z"} | value:2026-08-31T11:15:26Z | value:https://api.github.com/repos/huggingface/datasets/issues/8395/labels{/name} | value:8395 | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8394/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/8394 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8394/timeline | json:[] | value:5060755332 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8394 | value:IterableDataset.repeat() drops the Arrow fast path and widens numpy dtypes | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-04T13:01:08Z | 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/8394/events | value:False | value:2026-08-31T09:38:02Z | value:### Describe the bug
`RepeatExamplesIterable` is the only wrapper in `iterable_dataset.py` that does not forward `iter_arrow`, `is_typed` and `features` to the iterable it wraps. Two things follow from that: `.repeat()` drops the Arrow fast path, and under `numpy` format it silently widens the declared dtypes.
... | value:https://api.github.com/repos/huggingface/datasets/issues/8394/comments | value:I_kwDODunzps8AAAABLaT_hA | value:closed | value:0 | value:2026-08-31T09:38:02Z | value:https://api.github.com/repos/huggingface/datasets/issues/8394/labels{/name} | value:8394 | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | true | ||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8393/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/8393 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8393/timeline | json:[] | value:5057788029 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8393 | value:CPU Bottleneck in `IterableDataset.state_dict()` when shuffling multiple shards | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-04T07:16:30Z | 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/8393/events | value:False | value:2026-09-04T13:33:37Z | value:**Describe the bug**
There is a significant CPU bottleneck in `IterableDataset` dataloading when `buffer_size` and `max_buffer_input_shards` are greater than 1 (i.e. highly parallelized streaming with a shuffle buffer).
Currently, `_BaseExamplesIterable.state_dict()` falls back to performing a full recursive `co... | value:https://api.github.com/repos/huggingface/datasets/issues/8393/comments | value:I_kwDODunzps8AAAABLXe4fQ | value:closed | value:0 | value:2026-09-04T13:33:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8393/labels{/name} | value:8393 | json:{"login": "yuxin00j", "id": 255896926, "node_id": "U_kgDOD0CtXg", "avatar_url": "https://avatars.githubusercontent.com/u/255896926?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yuxin00j", "html_url": "https://github.com/yuxin00j", "followers_url": "https://api.github.com/users/yuxin00j/followers", ... | true | ||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8392/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/8392 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8392/timeline | json:[] | value:5057387644 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8392 | value:Add Spark Connect support for Spark datasets | value:2026-08-04T06:20:14Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8392/events | value:False | value:## What does this PR do?
Adds Spark Connect support to `Dataset.from_spark` and `IterableDataset.from_spark` by replacing client-side RDD and `SparkContext` calls with DataFrame APIs supported by Spark Connect.
- discovers partitions with `spark_partition_id()` for streaming datasets
- runs the shared-cache pro... | value:https://api.github.com/repos/huggingface/datasets/issues/8392/comments | value:PR_kwDODunzps76YcK8 | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8392", "html_url": "https://github.com/huggingface/datasets/pull/8392", "diff_url": "https://github.com/huggingface/datasets/pull/8392.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8392.patch", "merged_at": null} | value:2026-08-07T07:01:05Z | value:https://api.github.com/repos/huggingface/datasets/issues/8392/labels{/name} | value:8392 | json:{"login": "zhengruifeng", "id": 7322292, "node_id": "MDQ6VXNlcjczMjIyOTI=", "avatar_url": "https://avatars.githubusercontent.com/u/7322292?v=4", "gravatar_id": "", "url": "https://api.github.com/users/zhengruifeng", "html_url": "https://github.com/zhengruifeng", "followers_url": "https://api.github.com/users/zheng... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8391/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/8391 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8391/timeline | json:[] | value:5057110196 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8391 | value:Nanosecond timestamps are silently truncated even when Value("timestamp[ns]") is requested | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-04T05:36:19Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8391/events | value:False | value:### Describe the bug
`Dataset.from_dict` / `from_list` / `map` silently drop sub-microsecond precision, even when the user explicitly declares a nanosecond feature type. The resulting dataset reports `timestamp[ns]` but the data has already been truncated to microseconds.
### Steps to reproduce the bug
```pyth... | value:https://api.github.com/repos/huggingface/datasets/issues/8391/comments | value:I_kwDODunzps8AAAABLW1gtA | value:open | value:1 | value:2026-08-06T02:08:27Z | value:https://api.github.com/repos/huggingface/datasets/issues/8391/labels{/name} | value:8391 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8390/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/8390 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8390/timeline | json:[] | value:5057108509 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8390 | value:to_json cannot round-trip temporal columns: all timestamps written as epoch milliseconds | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-04T05:36:00Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8390/events | value:False | value:### Describe the bug
`Dataset.to_json()` serializes every temporal column as a bare epoch integer at a single fixed resolution (pandas' `date_unit="ms"` default), with no unit recorded in the output. Timestamps therefore do not survive a `to_json` → `from_json` round trip: they either come back silently wrong or... | value:https://api.github.com/repos/huggingface/datasets/issues/8390/comments | value:I_kwDODunzps8AAAABLW1aHQ | value:open | value:1 | value:2026-08-10T23:10:37Z | value:https://api.github.com/repos/huggingface/datasets/issues/8390/labels{/name} | value:8390 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8389/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/8389 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8389/timeline | json:[] | value:5057105324 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8389 | value:Keep the id of List and LargeList features when visiting them | value:2026-08-04T05:35:23Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8389/events | value:False | value:`_visit()` and `_visit_with_path()` rebuild `List` / `LargeList` in order to recurse into the inner feature, but they don't carry `id` over, so it is silently reset to `None`. Every other feature type is passed through untouched and keeps its `id`.
`Features.__init__` runs `_visit()` on every column (to convert ... | value:https://api.github.com/repos/huggingface/datasets/issues/8389/comments | value:PR_kwDODunzps76XicZ | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8389", "html_url": "https://github.com/huggingface/datasets/pull/8389", "diff_url": "https://github.com/huggingface/datasets/pull/8389.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8389.patch", "merged_at": null} | value:2026-08-04T05:35:23Z | value:https://api.github.com/repos/huggingface/datasets/issues/8389/labels{/name} | value:8389 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8388/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/8388 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8388/timeline | json:[] | value:5057105180 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8388 | value:Allow to_iterable_dataset() on a dataset with no rows | value:2026-08-04T05:35:21Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8388/events | value:False | value:`Dataset.to_iterable_dataset()` defaults to `num_shards=1`, but the guard rejects any `num_shards > len(self)`, so it always raises on an empty dataset:
```python
>>> Dataset.from_dict({"a": []}).to_iterable_dataset()
ValueError: Unable to shard a dataset of size 0 into 1 shards (the number of shards exceeds the... | value:https://api.github.com/repos/huggingface/datasets/issues/8388/comments | value:PR_kwDODunzps76Xiab | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8388", "html_url": "https://github.com/huggingface/datasets/pull/8388", "diff_url": "https://github.com/huggingface/datasets/pull/8388.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8388.patch", "merged_at": null} | value:2026-08-04T05:35:21Z | value:https://api.github.com/repos/huggingface/datasets/issues/8388/labels{/name} | value:8388 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8387/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/8387 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8387/timeline | json:[] | value:5057055820 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8387 | value:Fix `deepcopy` bottleneck in state_dict() | value:2026-08-04T05:26:28Z | 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/8387/events | value:False | value:2026-09-04T13:33:36Z | value:Fix #8393
This PR significantly resolves a major CPU bottleneck in `IterableDataset` dataloading when `buffer_size` and `max_buffer_input_shards` are > 1.
**The Bug:**
Currently, `_BaseExamplesIterable.state_dict()` falls back to performing a full recursive `copy.deepcopy(self._state_dict)`.
In nested it... | value:https://api.github.com/repos/huggingface/datasets/issues/8387/comments | value:PR_kwDODunzps76XYVI | value:closed | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8387", "html_url": "https://github.com/huggingface/datasets/pull/8387", "diff_url": "https://github.com/huggingface/datasets/pull/8387.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8387.patch", "merged_at": "2026-09-04T13:33:36Z"} | value:2026-09-04T13:33:36Z | value:https://api.github.com/repos/huggingface/datasets/issues/8387/labels{/name} | value:8387 | json:{"login": "yuxin00j", "id": 255896926, "node_id": "U_kgDOD0CtXg", "avatar_url": "https://avatars.githubusercontent.com/u/255896926?v=4", "gravatar_id": "", "url": "https://api.github.com/users/yuxin00j", "html_url": "https://github.com/yuxin00j", "followers_url": "https://api.github.com/users/yuxin00j/followers", ... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8386/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/8386 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8386/timeline | json:[] | value:5056333190 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8386 | value:map(batched=True, drop_last_batch=True) returns the input unchanged when the dataset is smaller than batch_size | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-04T03:21:59Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8386/events | value:False | value:### Describe the bug
When `drop_last_batch=True` and the dataset is smaller than `batch_size`, `Dataset.map(batched=True, ...)` returns the **input unchanged** instead of an empty dataset. The mapped function is never called, but the original rows are passed through with the original (unbatched) schema.
The ite... | value:https://api.github.com/repos/huggingface/datasets/issues/8386/comments | value:I_kwDODunzps8AAAABLWGFhg | value:open | value:1 | value:2026-09-18T16:59:48Z | value:https://api.github.com/repos/huggingface/datasets/issues/8386/labels{/name} | value:8386 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8385/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/8385 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8385/timeline | json:[] | value:5056320016 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8385 | value:Fix save_to_disk writing no shard for a dataset with no rows | value:2026-08-04T03:19:50Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8385/events | value:False | value:`save_to_disk` computes the number of shards from the estimated size and then clamps it to the number of rows (added in #7912):
```python
num_shards = min(len(self.data), num_shards)
```
For a dataset with no rows that gives `0` shards. No `.arrow` file is written, `state.json` gets an empty `"_data_files"`, an... | value:https://api.github.com/repos/huggingface/datasets/issues/8385/comments | value:PR_kwDODunzps76VCkY | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8385", "html_url": "https://github.com/huggingface/datasets/pull/8385", "diff_url": "https://github.com/huggingface/datasets/pull/8385.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8385.patch", "merged_at": null} | value:2026-08-04T03:19:50Z | value:https://api.github.com/repos/huggingface/datasets/issues/8385/labels{/name} | value:8385 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8384/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/8384 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8384/timeline | json:[] | value:5056316898 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8384 | value:Fix IterableDataset.remove_columns when a single column name is passed | value:2026-08-04T03:19:23Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8384/events | value:False | value:`IterableDataset.remove_columns` accepts either a single column name or a list, but it only ever tests `col in column_names`. When a string is passed, that becomes a substring test, so any existing column whose name is contained in the one being removed is dropped from `info.features`:
```python
ds = Dataset.fro... | value:https://api.github.com/repos/huggingface/datasets/issues/8384/comments | value:PR_kwDODunzps76VB6Y | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8384", "html_url": "https://github.com/huggingface/datasets/pull/8384", "diff_url": "https://github.com/huggingface/datasets/pull/8384.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8384.patch", "merged_at": null} | value:2026-08-04T03:19:23Z | value:https://api.github.com/repos/huggingface/datasets/issues/8384/labels{/name} | value:8384 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8383/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/8383 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8383/timeline | json:[] | value:5056314094 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8383 | value:Apply the indices mapping in Dataset.unique | value:2026-08-04T03:18:59Z | value:CONTRIBUTOR | value:https://api.github.com/repos/huggingface/datasets/issues/8383/events | value:False | value:`Dataset.unique` only takes the indices mapping into account when it has a different number of rows than the underlying table:
```python
if self._indices is not None and self._indices.num_rows != self._data.num_rows:
```
An indices mapping can also repeat some rows and drop others while keeping the same length.... | value:https://api.github.com/repos/huggingface/datasets/issues/8383/comments | value:PR_kwDODunzps76VBUb | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8383", "html_url": "https://github.com/huggingface/datasets/pull/8383", "diff_url": "https://github.com/huggingface/datasets/pull/8383.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8383.patch", "merged_at": null} | value:2026-08-04T03:18:59Z | value:https://api.github.com/repos/huggingface/datasets/issues/8383/labels{/name} | value:8383 | json:{"login": "LeSingh1", "id": 197028264, "node_id": "U_kgDOC75pqA", "avatar_url": "https://avatars.githubusercontent.com/u/197028264?v=4", "gravatar_id": "", "url": "https://api.github.com/users/LeSingh1", "html_url": "https://github.com/LeSingh1", "followers_url": "https://api.github.com/users/LeSingh1/followers", ... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8382/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/8382 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8382/timeline | json:[] | value:5053067850 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8382 | value:Fix IterableDataset.to_dict() and to_polars() returning an empty generator | value:2026-08-03T18:35:31Z | value:CONTRIBUTOR | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | value:https://api.github.com/repos/huggingface/datasets/issues/8382/events | value:False | value:2026-09-02T18:47:11Z | value:### The bug
`IterableDataset.to_dict()` and `IterableDataset.to_polars()` each contain a bare `yield` in their `batched=True` branch. A bare `yield` anywhere in a function body makes the whole function a generator function, so both methods return a generator on every call, including the `batched=False` default. ... | value:https://api.github.com/repos/huggingface/datasets/issues/8382/comments | value:PR_kwDODunzps76KnWs | value:closed | value:4 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8382", "html_url": "https://github.com/huggingface/datasets/pull/8382", "diff_url": "https://github.com/huggingface/datasets/pull/8382.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8382.patch", "merged_at": null} | value:2026-09-02T18:47:11Z | value:https://api.github.com/repos/huggingface/datasets/issues/8382/labels{/name} | value:8382 | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8381/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/8381 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8381/timeline | json:[] | value:5053065106 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8381 | value:IterableDataset.to_dict() and to_polars() return an empty generator instead of the data | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-08-03T18:35:13Z | 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/8381/events | value:False | value:2026-10-05T14:18:58Z | value:### Describe the bug
`IterableDataset.to_dict()` and `IterableDataset.to_polars()` each have a bare `yield` in their `batched=True` branch, which makes the whole method a generator function. Calling either one the way its docstring shows returns a generator that yields nothing, and the value the `else` branch re... | value:https://api.github.com/repos/huggingface/datasets/issues/8381/comments | value:I_kwDODunzps8AAAABLS-nkg | value:closed | value:3 | value:2026-10-05T14:18:58Z | value:https://api.github.com/repos/huggingface/datasets/issues/8381/labels{/name} | value:8381 | json:{"login": "ebarkhordar", "id": 22658149, "node_id": "MDQ6VXNlcjIyNjU4MTQ5", "avatar_url": "https://avatars.githubusercontent.com/u/22658149?v=4", "gravatar_id": "", "url": "https://api.github.com/users/ebarkhordar", "html_url": "https://github.com/ebarkhordar", "followers_url": "https://api.github.com/users/ebarkh... | true | ||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8380/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/8380 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8380/timeline | json:[] | value:5035622149 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8380 | value:perf(formatting): Python row iteration -- up to 1.43x faster | value:2026-08-01T06:13:11Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8380/events | value:False | value:## TL;DR
- Read standard Arrow cells directly and reuse the stateless Python extractor, while retaining ArrayXD conversion.
- This speeds up default Python `for row in ds` iteration and `ds[i]` access; `ds.map()` is unchanged.
- The recorded benchmark measures iteration time on a single core with stock defaults,... | value:https://api.github.com/repos/huggingface/datasets/issues/8380/comments | value:PR_kwDODunzps75T9aA | value:open | value:1 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8380", "html_url": "https://github.com/huggingface/datasets/pull/8380", "diff_url": "https://github.com/huggingface/datasets/pull/8380.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8380.patch", "merged_at": null} | value:2026-10-23T09:10:20Z | value:https://api.github.com/repos/huggingface/datasets/issues/8380/labels{/name} | value:8380 | json:{"login": "dexhunter", "id": 6930518, "node_id": "MDQ6VXNlcjY5MzA1MTg=", "avatar_url": "https://avatars.githubusercontent.com/u/6930518?v=4", "gravatar_id": "", "url": "https://api.github.com/users/dexhunter", "html_url": "https://github.com/dexhunter", "followers_url": "https://api.github.com/users/dexhunter/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8378/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/8378 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8378/timeline | json:[] | value:5029934393 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8378 | value:feat: add return_file_name option to JSON loader | value:2026-07-31T13:19:19Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8378/events | value:False | value:## What does this PR do?
Adds a `return_file_name` option to the JSON loader. When enabled, the resulting dataset includes a `file_name` column indicating the source file of each row.
## Motivation
When training is interrupted and resumed, it's useful to know which file each row came from so already-trai... | value:https://api.github.com/repos/huggingface/datasets/issues/8378/comments | value:PR_kwDODunzps75Bwbf | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8378", "html_url": "https://github.com/huggingface/datasets/pull/8378", "diff_url": "https://github.com/huggingface/datasets/pull/8378.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8378.patch", "merged_at": null} | value:2026-07-31T13:19:19Z | value:https://api.github.com/repos/huggingface/datasets/issues/8378/labels{/name} | value:8378 | json:{"login": "RenhaoSong", "id": 305858618, "node_id": "U_kgDOEjsIOg", "avatar_url": "https://avatars.githubusercontent.com/u/305858618?v=4", "gravatar_id": "", "url": "https://api.github.com/users/RenhaoSong", "html_url": "https://github.com/RenhaoSong", "followers_url": "https://api.github.com/users/RenhaoSong/foll... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8377/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/8377 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8377/timeline | json:[] | value:5027590302 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8377 | value:Fix PandasArrayExtensionDtype._metadata: must be a tuple, not a string | value:2026-07-31T07:47:13Z | value:NONE | json:{"login": "saket3395", "id": 55333527, "node_id": "MDQ6VXNlcjU1MzMzNTI3", "avatar_url": "https://avatars.githubusercontent.com/u/55333527?v=4", "gravatar_id": "", "url": "https://api.github.com/users/saket3395", "html_url": "https://github.com/saket3395", "followers_url": "https://api.github.com/users/saket3395/fo... | value:https://api.github.com/repos/huggingface/datasets/issues/8377/events | value:False | value:2026-08-02T15:55:12Z | value:## Summary
Fixes #8375 — `PandasArrayExtensionDtype._metadata` is a plain string instead of a tuple, causing `AttributeError: 'PandasArrayExtensionDtype' object has no attribute 'v'` whenever pandas compares two instances of this dtype (e.g. during a boolean-mask filter on a DataFrame with an array-feature column... | value:https://api.github.com/repos/huggingface/datasets/issues/8377/comments | value:PR_kwDODunzps746K85 | value:closed | value:2 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8377", "html_url": "https://github.com/huggingface/datasets/pull/8377", "diff_url": "https://github.com/huggingface/datasets/pull/8377.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8377.patch", "merged_at": null} | value:2026-08-02T15:55:12Z | value:https://api.github.com/repos/huggingface/datasets/issues/8377/labels{/name} | value:8377 | json:{"login": "saket3395", "id": 55333527, "node_id": "MDQ6VXNlcjU1MzMzNTI3", "avatar_url": "https://avatars.githubusercontent.com/u/55333527?v=4", "gravatar_id": "", "url": "https://api.github.com/users/saket3395", "html_url": "https://github.com/saket3395", "followers_url": "https://api.github.com/users/saket3395/fo... | value:False | true | |||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8376/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/8376 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8376/timeline | json:[] | value:5025670677 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8376 | value:Fix PandasArrayExtensionDtype._metadata to use a tuple | value:2026-07-31T02:02:46Z | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8376/events | value:False | value:## What was wrong
`PandasArrayExtensionDtype` set `_metadata = "value_type"` (a string), but pandas expects a tuple of attribute names, e.g. `("value_type",)`.
When pandas compares dtypes (during things like `df[mask]`), it loops over `_metadata` and looks up each entry as an attribute. With a string, it iterat... | value:https://api.github.com/repos/huggingface/datasets/issues/8376/comments | value:PR_kwDODunzps740EFW | value:open | value:0 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8376", "html_url": "https://github.com/huggingface/datasets/pull/8376", "diff_url": "https://github.com/huggingface/datasets/pull/8376.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8376.patch", "merged_at": null} | value:2026-07-31T06:09:30Z | value:https://api.github.com/repos/huggingface/datasets/issues/8376/labels{/name} | value:8376 | json:{"login": "kohankhaki", "id": 56772198, "node_id": "MDQ6VXNlcjU2NzcyMTk4", "avatar_url": "https://avatars.githubusercontent.com/u/56772198?v=4", "gravatar_id": "", "url": "https://api.github.com/users/kohankhaki", "html_url": "https://github.com/kohankhaki", "followers_url": "https://api.github.com/users/kohankhak... | value:False | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8375/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/8375 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8375/timeline | json:[] | value:5025604085 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8375 | value:`PandasArrayExtensionDtype._metadata` should be a tuple, not a string | json:{"blocked_by": 0, "total_blocked_by": 0, "blocking": 0, "total_blocking": 0} | value:2026-07-31T01:50:02Z | json:{"total": 0, "completed": 0, "percent_completed": 0} | value:NONE | value:https://api.github.com/repos/huggingface/datasets/issues/8375/events | value:False | value:### Describe the bug
There's a small typo in `PandasArrayExtensionDtype` (`src/datasets/features/features.py`):
```python
_metadata = "value_type"
```
Pandas expects `_metadata` to be a tuple of attribute names, like `("value_type",)`, not a plain string ([API documentation ref](https://pandas.pydata.org/docs... | value:https://api.github.com/repos/huggingface/datasets/issues/8375/comments | value:I_kwDODunzps8AAAABK4yh9Q | value:open | value:1 | value:2026-08-12T06:30:29Z | value:https://api.github.com/repos/huggingface/datasets/issues/8375/labels{/name} | value:8375 | json:{"login": "kohankhaki", "id": 56772198, "node_id": "MDQ6VXNlcjU2NzcyMTk4", "avatar_url": "https://avatars.githubusercontent.com/u/56772198?v=4", "gravatar_id": "", "url": "https://api.github.com/users/kohankhaki", "html_url": "https://github.com/kohankhaki", "followers_url": "https://api.github.com/users/kohankhak... | true | |||||||||||
json:{"url": "https://api.github.com/repos/huggingface/datasets/issues/8374/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/8374 | json:[] | value:https://api.github.com/repos/huggingface/datasets/issues/8374/timeline | json:[] | value:5020005801 | value:https://api.github.com/repos/huggingface/datasets | value:https://api.github.com/repos/huggingface/datasets/issues/8374 | value:feat: Qdrant Search Index | value:2026-07-30T13:04:05Z | value:NONE | json:{"login": "anush008", "id": 46051506, "node_id": "MDQ6VXNlcjQ2MDUxNTA2", "avatar_url": "https://avatars.githubusercontent.com/u/46051506?v=4", "gravatar_id": "", "url": "https://api.github.com/users/anush008", "html_url": "https://github.com/anush008", "followers_url": "https://api.github.com/users/anush008/follow... | value:https://api.github.com/repos/huggingface/datasets/issues/8374/events | value:False | value:2026-10-07T17:37:42Z | value:Resolves #5698
## Description
This PR adds Qdrant as a vector search backend alongside FAISS and Elasticsearch. Associated documentation and tests are also included. The Qdrant client dependency is marked optional and installed on demand.
[Qdrant](https://qdrant.tech/) is an open-source vector search en... | value:https://api.github.com/repos/huggingface/datasets/issues/8374/comments | value:PR_kwDODunzps74hvA5 | value:closed | value:4 | json:{"url": "https://api.github.com/repos/huggingface/datasets/pulls/8374", "html_url": "https://github.com/huggingface/datasets/pull/8374", "diff_url": "https://github.com/huggingface/datasets/pull/8374.diff", "patch_url": "https://github.com/huggingface/datasets/pull/8374.patch", "merged_at": null} | value:2026-10-07T17:37:43Z | value:https://api.github.com/repos/huggingface/datasets/issues/8374/labels{/name} | value:8374 | json:{"login": "anush008", "id": 46051506, "node_id": "MDQ6VXNlcjQ2MDUxNTA2", "avatar_url": "https://avatars.githubusercontent.com/u/46051506?v=4", "gravatar_id": "", "url": "https://api.github.com/users/anush008", "html_url": "https://github.com/anush008", "followers_url": "https://api.github.com/users/anush008/follow... | value:False | true |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.