anhaidgroup/polaris-wikitables-v1

WikiTables v1

WikiTables is one of six datasets in [Polaris: Learning to Generate Table Descriptions from Retrieval Feedback](https://arxiv.org/abs/2608.17171), alongside aw, arctic, lter, ecir, and wtr.

It holds 3,361 tables scraped from Wikipedia articles and 57 keyword queries over them. For each query–table pair, a person scored how well that table answers that query; those scores are the relevance judgments, and they live in qrels.csv.

The tables have no names. What describes a table is its column names and the page context around it — the article title, the section heading, and the caption.

Each Polaris dataset comes in two versions. v1, this one, has the table metadata, the queries, and the relevance judgments. v2 adds the tuples — the contents of each table — and is at polaris-wikitables-v2.

Files

`queries.csv` — one row per query. Note that query_id is a plain integer here, not the q1 form used by aw, arctic, lter, and ecir.

query_id,query
1,world interest rates table
2,2008 beijing olympics

`qrels.csv` — one row per query–table pair that a person scored, graded 0, 1, or 2.

query_id,table_id,relevance_score
1,table-0370-614,2

Judgments came with the source benchmark rather than being made here, and they cover a candidate pool larger than the 3,361 tables released, so 1,841 rows point at tables that are not in metadata.csv. All 1,841 are grade 0, so every gold table is present. Drop them if your loader expects every table_id to resolve.

`metadata.csv` — one row per table, here the table above.

table_id,column_names,table_context
table-0370-614,"[""Date"", ""Deposit facility"", ""Main refinancing operations"", ""Marginal lending facility""]","{""pgTitle"": ""Eurozone"", ""secondTitle"": ""Interest rates"", ""caption"": ""Interest rates""}"

Statistics

A table counts as gold if it scores above 0 for at least one query.

Download

The three files come to about 0.96 MB. You do not need a Hugging Face account to download them.

Option 1 — click the Files tab at the top of this page and save each file.

Option 2 — command line (recommended):

pip install huggingface_hub
hf download anhaidgroup/polaris-wikitables-v1 --repo-type dataset --local-dir wikitables

Polaris has six datasets in total and WikiTables is one of them. Each sits in its own repository, so to download all six quickly — the v1 repositories, without tuples:

for d in aw arctic lter ecir wikitables wtr; do
  hf download anhaidgroup/polaris-$d-v1 --repo-type dataset --local-dir polaris_v1/$d
done

Usage

column_names and table_context are an array and an object, so they need parsing when you load the file. For example:

import ast
import pandas as pd

metadata = pd.read_csv("wikitables/metadata.csv")
queries = pd.read_csv("wikitables/queries.csv", dtype={"query_id": str})
qrels = pd.read_csv("wikitables/qrels.csv", dtype={"query_id": str})

metadata["columns"] = metadata["column_names"].apply(ast.literal_eval)
metadata["context"] = metadata["table_context"].apply(ast.literal_eval)

relevant = qrels.loc[(qrels.query_id == "1") & (qrels.relevance_score > 0), "table_id"].tolist()

The dtype={"query_id": str} keeps the ids as strings, so the same code works across all six datasets.

How is this dataset created?

The tables come from a corpus of Wikipedia tables that Bhagavatula et al. (ISWC 2015) assembled for entity linking. The retrieval benchmark built over it pairs 60 queries with 1.6 million tables and scores each pair 0, 1, or 2: irrelevant, partially relevant, fully relevant. A Wikipedia table has no title of its own, so what identifies it here is its column names and the page it sits on.

Producing a description for 1.6 million tables was not practical, so the Polaris authors cut the corpus down. For each query they ran BM25 over the column names and page context, took the 50 highest-scoring tables, and then added back every table already scored above 0.

That leaves the corpus harder than a random sample would, not easier. What survives is every relevant table plus the non-relevant tables that BM25 ranks highest — the ones a system is most likely to confuse them with. Queries left without a single relevant table were dropped, giving 3,361 tables and 57 queries.

Citation

@misc{cai2026polaris,
  title         = {Polaris: Learning to Generate Table Descriptions from Retrieval Feedback},
  author        = {Cai, Ting and Phan, Tuan Minh and Doan, AnHai},
  year          = {2026},
  eprint        = {2608.17171},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2608.17171},
  url           = {https://arxiv.org/abs/2608.17171}
}

Please also cite the corpus and the benchmark it is built on:

@inproceedings{bhagavatula2015tabel,
  author    = {Chandra Sekhar Bhagavatula and Thanapon Noraset and Doug Downey},
  title     = {{TabEL}: Entity Linking in Web Tables},
  booktitle = {Proceedings of the International Semantic Web Conference ({ISWC})},
  pages     = {425--441},
  year      = {2015},
}

@inproceedings{zhang2018adhoc,
  author    = {Shuo Zhang and Krisztian Balog},
  title     = {Ad Hoc Table Retrieval Using Semantic Similarity},
  booktitle = {Proceedings of the Web Conference ({WWW})},
  pages     = {1553--1562},
  year      = {2018},
}

License

Contact

Email minhrua@cs.wisc.edu, valid until May 2029. After that, email anhai@cs.wisc.edu.