RelativeDB/rt-j-int4
Tabular Classification • Updated
name stringclasses 7
values | domain stringclasses 7
values | description stringclasses 7
values | num_tables int64 3 15 | num_rows int64 97.6k 44.8M | num_cols int64 15 136 | num_tasks int64 4 9 | tasks_binary_classification int64 1 4 | tasks_regression int64 1 3 | tasks_recommendation int64 0 3 | start_timestamp stringclasses 7
values | val_timestamp stringdate 2005-01-01 00:00:00 2020-10-01 00:00:00 | test_timestamp stringdate 2010-01-01 00:00:00 2021-01-01 00:00:00 | size_gb float64 0 7.48 | license stringclasses 6
values | source_url stringclasses 7
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
rel-amazon | E-commerce (product reviews) | Amazon product reviews: customers, products, and time-stamped reviews and ratings across the Amazon catalog. | 3 | 23,218,245 | 15 | 8 | 2 | 3 | 3 | 2008-01-01 00:00:00 | 2015-10-01 00:00:00 | 2016-01-01 00:00:00 | 7.4787 | No explicit license (McAuley Lab; citation requested) | https://cseweb.ucsd.edu/~jmcauley/datasets/amazon_v2/ |
rel-avito | Online advertising | Avito online classifieds: users, ads, search queries, and impression / click / visit streams. | 8 | 24,653,915 | 42 | 6 | 4 | 1 | 1 | 2015-04-25 00:00:00 | 2015-05-08 00:00:00 | 2015-05-14 00:00:00 | 0.4886 | Kaggle competition rules: competition use only, no redistribution | https://www.kaggle.com/c/avito-context-ad-clicks/rules |
rel-event | Social events / recommendation | Event recommendation: users, events, attendance records, and social and interest signals. | 5 | 44,822,992 | 129 | 6 | 4 | 2 | 0 | 1912-01-01 05:59:24.003000 | 2012-11-21 00:00:00 | 2012-11-29 00:00:00 | 0.1199 | Kaggle competition rules: competition use only, no redistribution | https://www.kaggle.com/c/event-recommendation-engine-challenge/rules |
rel-f1 | Motorsport (Formula 1) | Formula 1 motorsport database: races, drivers, constructors, circuits, race results, qualifying, and championship standings. | 9 | 97,606 | 67 | 6 | 2 | 3 | 1 | 1950-05-13 00:00:00 | 2005-01-01 00:00:00 | 2010-01-01 00:00:00 | 0.0014 | CC BY-NC-SA 3.0 (Ergast database images) | https://ergast.com/mrd/db/ |
rel-hm | Fashion retail | H&M e-commerce: customers, articles, and time-stamped purchase transactions. | 3 | 16,931,173 | 37 | 4 | 1 | 2 | 1 | 2019-09-07 00:00:00 | 2020-09-07 00:00:00 | 2020-09-14 00:00:00 | 0.276 | Kaggle competition rules: non-commercial and academic research only, no redistribution | https://www.kaggle.com/competitions/h-and-m-personalized-fashion-recommendations/rules |
rel-stack | Online Q&A (Stack Exchange) | Stack Exchange Q&A: users, posts, comments, votes, badges, and post links. | 7 | 5,399,818 | 50 | 5 | 2 | 1 | 2 | 2009-02-02 00:00:00 | 2020-10-01 00:00:00 | 2021-01-01 00:00:00 | 1.0739 | CC BY-SA 4.0 (older posts CC BY-SA 2.5/3.0) | https://archive.org/details/stackexchange |
rel-trial | Clinical trials / healthcare | ClinicalTrials.gov clinical trials: studies, outcomes, adverse events, eligibilities, sponsors, conditions, and facilities. | 15 | 5,852,157 | 136 | 9 | 4 | 3 | 2 | 2000-01-01 00:00:00 | 2020-01-01 00:00:00 | 2021-01-01 00:00:00 | 0.6827 | ClinicalTrials.gov terms (US Government data; attribution required) | https://clinicaltrials.gov/about-site/terms-conditions |
RelBench v1
The RelBench v1 databases and tasks, one database per subdirectory. The v2-only databases and tasks are in stanford-star/relbench-v2-extra. regression_stds.json holds the train-split target standard deviations that turn MAE into NMAE.
<dataset>/
manifest.yaml # tables, primary keys, foreign-key graph, val/test timestamps
schema.svg # ER diagram
db/*.parquet # relational tables
tasks/<task>/manifest.yaml # task spec
tasks/<task>/{train,val,test}.parquet
import relbench
ds = relbench.load_dataset("rel-f1")
task = ds.load_task("driver-position")
License
Each database keeps the license of its upstream source, listed in the license and source_url columns of STATS/databases.parquet. Our contributions (the conversion to this format, schemas, task definitions, splits and statistics) are CDLA-Permissive-2.0.
Citation
@inproceedings{relbench,
title={RelBench: A Benchmark for Deep Learning on Relational Databases},
author={Robinson, Joshua and Ranjan, Rishabh and Hu, Weihua and Huang, Kexin and Han, Jiaqi and Dobles, Alejandro and Fey, Matthias and Lenssen, Jan Eric and Yuan, Yiwen and Zhang, Zecheng and He, Xinwei and Leskovec, Jure},
booktitle={Advances in Neural Information Processing Systems},
year={2024}
}
- Downloads last month
- 54,108