I'll systematically check for data quality issues across the different types of columns in this table tennis dataset. {"action": "tool_call", "reasoning": "First, I'll check for logical consistency in the match results - sets won should align with game scores, and there should be a winner in each match. I'll also check if the sets add up correctly and if HomeWinner aligns with who actually won.", "tool": "query_sql", "args": {"sql": "SELECT COUNT(*) as total_matches, COUNT(CASE WHEN Sets_P1 + Sets_P2 = 0 THEN 1 END) as no_sets_played, COUNT(CASE WHEN Sets_P1 = Sets_P2 THEN 1 END) as tied_matches, COUNT(CASE WHEN Sets_P1 > 3 OR Sets_P2 > 3 THEN 1 END) as impossible_set_counts, COUNT(CASE WHEN (Sets_P1 >= 3 AND HomeWinner = 0) OR (Sets_P2 >= 3 AND HomeWinner = 1) THEN 1 END) as winner_mismatch, COUNT(CASE WHEN Sets_P1 + Sets_P2 > 5 THEN 1 END) as too_many_sets FROM raw.table_tennis_sept2022", "purpose": "Check for logical inconsistencies in match outcomes - ties, impossible set counts, winner mismatches", "source": "raw"}}