vaishali/multitabqa-base-sql
Table Question Answering
•
Updated
•
9
•
4
tables
sequence | table_names
sequence | query
string | answer
string | db_name
string | source
string | target
string |
---|---|---|---|---|---|---|
["{\"columns\":[\"player_id\",\"year\",\"game_num\",\"game_id\",\"team_id\",\"league_id\",\"gp\",\"s(...TRUNCATED) | [
"all_star",
"manager_award_vote"
] | "SELECT T1.game_num, AVG(T2.votes_first) FROM all_star as T1 JOIN manager_award_vote as T2 ON T1.lea(...TRUNCATED) | "{\"columns\":[\"game_num\",\"AVG(T2.votes_first)\"],\"index\":[0],\"data\":[[\"0\",\"4.5427945763\"(...TRUNCATED) | null | null | null |
["{\"columns\":[\"player_id\",\"year\",\"game_num\",\"game_id\",\"team_id\",\"league_id\",\"gp\",\"s(...TRUNCATED) | [
"all_star",
"player_college"
] | "SELECT MAX(T1.starting_pos), T2.year FROM all_star as T1 JOIN player_college as T2 ON T1.player_id (...TRUNCATED) | {"columns":["MAX(T1.starting_pos)","year"],"index":[0],"data":[["","1927"]]} | null | null | null |
["{\"columns\":[\"player_id\",\"award_id\",\"year\",\"league_id\",\"tie\",\"notes\"],\"index\":[0,1,(...TRUNCATED) | [
"manager_award",
"team_half"
] | "SELECT league_id, year FROM manager_award WHERE player_id = \"wedgeer01\" AND award_id = \"TSN Mana(...TRUNCATED) | "{\"columns\":[\"league_id\",\"year\"],\"index\":[0,1,2,3],\"data\":[[\"AL\",\"1981\"],[\"AL\",\"201(...TRUNCATED) | null | null | null |
["{\"columns\":[\"player_id\",\"year\",\"game_num\",\"game_id\",\"team_id\",\"league_id\",\"gp\",\"s(...TRUNCATED) | [
"all_star",
"fielding"
] | "SELECT T1.starting_pos, SUM(T2.a) FROM all_star as T1 JOIN fielding as T2 ON T1.league_id = T2.leag(...TRUNCATED) | {"columns":["starting_pos","SUM(T2.a)"],"index":[0],"data":[["1","11686740233"]]} | null | null | null |
["{\"columns\":[\"player_id\",\"year\",\"game_num\",\"game_id\",\"team_id\",\"league_id\",\"gp\",\"s(...TRUNCATED) | [
"all_star",
"pitching_postseason"
] | "SELECT MAX(T1.year), T2.g_idp FROM all_star as T1 JOIN pitching_postseason as T2 ON T1.team_id = T2(...TRUNCATED) | {"columns":["MAX(T1.year)","g_idp"],"index":[0],"data":[["2015","0"]]} | null | null | null |
["{\"columns\":[\"customer_id\",\"datetime_payment\",\"payment_method_code\",\"amount_payment\"],\"i(...TRUNCATED) | [
"Customer_Payments",
"Customers"
] | "SELECT T1.amount_payment, T2.amount_outstanding FROM Customer_Payments as T1 JOIN Customers as T2 O(...TRUNCATED) | "{\"columns\":[\"amount_payment\",\"amount_outstanding\"],\"index\":[0,1,2,3,4,5,6,7,8],\"data\":[[\(...TRUNCATED) | null | null | null |
["{\"columns\":[\"Driver_ID\",\"Name\",\"Party\",\"Home_city\",\"Age\"],\"index\":[0,1,2,3,4,5,6,7,8(...TRUNCATED) | [
"driver",
"school_bus"
] | "SELECT Driver_ID FROM driver GROUP BY Driver_ID HAVING Driver_ID = 10 UNION SELECT Driver_ID FROM s(...TRUNCATED) | "{\"columns\":[\"Driver_ID\"],\"index\":[0,1,2,3,4],\"data\":[[\"3\"],[\"4\"],[\"7\"],[\"9\"],[\"10\(...TRUNCATED) | null | null | null |
["{\"columns\":[\"Branch_ID\",\"Name\",\"Open_year\",\"Address_road\",\"City\",\"membership_amount\"(...TRUNCATED) | [
"branch",
"member"
] | "SELECT MIN(Name) FROM branch WHERE Open_year = \"2013\" OR Address_road <> \"Cecilia Avenue\" AND m(...TRUNCATED) | {"columns":["MIN(Name)"],"index":[0],"data":[["Alexandre"]]} | null | null | null |
["{\"columns\":[\"player_id\",\"year\",\"game_num\",\"game_id\",\"team_id\",\"league_id\",\"gp\",\"s(...TRUNCATED) | [
"all_star",
"manager_half"
] | "SELECT AVG(T1.game_num), T2.year FROM all_star as T1 JOIN manager_half as T2 ON T1.league_id = T2.l(...TRUNCATED) | {"columns":["AVG(T1.game_num)","year"],"index":[0],"data":[["0.1361423985","1981"]]} | null | null | null |
["{\"columns\":[\"property_id\",\"property_type_code\",\"date_on_market\",\"date_sold\",\"property_n(...TRUNCATED) | [
"Properties",
"Properties"
] | "SELECT MIN(oth_feature_3), AVG(hse_feature_3), date_on_market, AVG(apt_feature_1), MIN(property_add(...TRUNCATED) | "{\"columns\":[\"MIN(oth_feature_3)\",\"AVG(hse_feature_3)\",\"date_on_market\",\"AVG(apt_feature_1)(...TRUNCATED) | null | null | null |
import pandas as pd
from datasets import load_dataset
multitableQA_pretraining = load_dataset("vaishali/multitabqa_pretraining")
for sample in multitableQA_pretraining['train']:
sql_query = sample['query']
input_table_names = sample["table_names"]
input_tables = [pd.read_json(table, orient='split') for table in sample['tables']]
answer = pd.read_json(sample['answer'], orient='split')
# flattened input/output
input_to_model = sample["source"]
target = sample["target"]
@inproceedings{pal-etal-2023-multitabqa,
title = "{M}ulti{T}ab{QA}: Generating Tabular Answers for Multi-Table Question Answering",
author = "Pal, Vaishali and
Yates, Andrew and
Kanoulas, Evangelos and
de Rijke, Maarten",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.348",
doi = "10.18653/v1/2023.acl-long.348",
pages = "6322--6334",
abstract = "Recent advances in tabular question answering (QA) with large language models are constrained in their coverage and only answer questions over a single table. However, real-world queries are complex in nature, often over multiple tables in a relational database or web page. Single table questions do not involve common table operations such as set operations, Cartesian products (joins), or nested queries. Furthermore, multi-table operations often result in a tabular output, which necessitates table generation capabilities of tabular QA models. To fill this gap, we propose a new task of answering questions over multiple tables. Our model, MultiTabQA, not only answers questions over multiple tables, but also generalizes to generate tabular answers. To enable effective training, we build a pre-training dataset comprising of 132,645 SQL queries and tabular answers. Further, we evaluate the generated tables by introducing table-specific metrics of varying strictness assessing various levels of granularity of the table structure. MultiTabQA outperforms state-of-the-art single table QA models adapted to a multi-table QA setting by finetuning on three datasets: Spider, Atis and GeoQuery.",
}