# salesforce/WikiSQL

A large annotated semantic parsing corpus for developing natural language interfaces.

Repository: https://github.com/salesforce/WikiSQL
Canonical: https://ross.abutalabs.com/products/wikisql
Language: HTML
License: BSD-3-Clause
License Family: permissive
Topics: natural-language, dataset, database, machine-learning, natural-language-processing, natural-language-interface
Archived: true
Last push: 2025-10-06T11:06:57+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 45, release rhythm 8, longevity 100
- inputs: {"age_days": 3309, "days_push": 331, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1805, forks 333 (observed 2026-08-28T04:05:38.989225+00:00)

## What it is
WikiSQL is a large crowd-sourced annotated dataset for semantic parsing, pairing natural language questions with SQL queries over Wikipedia tables. It was released alongside the Seq2SQL paper and serves as a standard benchmark for developing natural language interfaces to relational databases.

## Use cases
- train a text-to-SQL model
- benchmark semantic parsing accuracy on tables
- build a natural language interface for relational databases
- evaluate question answering over tabular data
- compare weakly supervised and supervised parsing approaches

## When to choose
- you need a large annotated corpus for text-to-SQL research
- you want a standard benchmark with a public leaderboard
- you are studying natural language interfaces to databases

## When to avoid
- you need a production natural language interface rather than a research dataset
- you need multi-table or complex SQL joins, since WikiSQL covers single-table queries
- you need actively maintained tooling, as the tokenizer dependencies are deprecated

## Facets
- artifact type: dataset
- maturity: maintenance
- function: nlp, machine-learning, parser, database
- domain: machine-learning, databases, data-science
- platform: python, cross-platform
- tags: semantic-parsing, text-to-sql, benchmark-dataset, question-answering, seq2sql, natural-language-processing

## Member repositories
- salesforce/WikiSQL (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:38.989225+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:21:47.329509+00:00, confidence not recorded.
  - readme: https://github.com/salesforce/WikiSQL (fetched 2026-08-28T04:05:38.989225+00:00, sha 1f8c077c58ab)
- Data as of 2026-08-30T08:39:29.467469+00:00.