I may find two ground truth SQL queries in dev dataset that may be wrong. This is from the data/bird/dev.json. These two SQL queries are marked as ground truth but they cannot be executed against the sqlite DB.
The first SQL query is
"SQL": "SELECT T3.district, CAST((T3.A13 - T3.A12) AS REAL) * 100 / T3.A12 FROM loan AS T1 INNER JOIN account AS T2 ON T1.account_id = T2.account_id INNER JOIN district AS T3 ON T2.district_id = T3.district_id WHERE T1.status = 'D'"
However it might be
"SQL": "SELECT T3.A2, CAST((T3.A13 - T3.A12) AS REAL) * 100 / T3.A12 FROM loan AS T1 INNER JOIN account AS T2 ON T1.account_id = T2.account_id INNER JOIN district AS T3 ON T2.district_id = T3.district_id WHERE T1.status = 'D'"
This fix is to replace T3.district with T3.A2. There is no district in T3 (which is district table). Per the table schema I think it is actually T3.A2.
The second SQL. query is
"SQL": "SELECT COUNT(driverId) FROM target WHERE raceId = 18"
"SQL": "SELECT COUNT(driverId) FROM results WHERE raceId = 18"
This fix is to replace target table which does not exist. I think the table should be results.
Hi team,
I may find two ground truth SQL queries in dev dataset that may be wrong. This is from the
data/bird/dev.json
. These two SQL queries are marked as ground truth but they cannot be executed against the sqlite DB.The first SQL query is
However it might be
This fix is to replace
T3.district
withT3.A2
. There is nodistrict
inT3
(which isdistrict
table). Per the table schema I think it is actuallyT3.A2
.The second SQL. query is
This fix is to replace
target
table which does not exist. I think the table should beresults
.