Each item is a text-to-sql correctness against live database example providing Task / question, Schema (DDL), Fixture data (INSERT statements), SQL query, Query category, Expected result. Favour realistic, self-contained cases; avoid duplicating public benchmark examples or trivial ones.
Secured after final acceptance. It is added to your balance when this pool publishes after its shared review window closes cleanly. Rejected items do not qualify.
Platform-authored spec, open on delivery.
Measured pipeline stats for this dataset. A dash means the platform does not publish that measure for this pool.
Approved public samples for this SQL Query Correctness (Text-to-SQL / Execution Accuracy) dataset. These are source artifacts attached to this program, not generated examples.