{"source_split": "dev", "source_row": 4443, "table_id": "2-17382981-2", "question": "What is the lowest game number on 20 July 2008?", "headers": ["Game", "Date", "Opponent", "Venue", "Result", "Attendance"], "types": ["real", "text", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[1, 0, "20 july 2008"]], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Game\"\ncol1 (text): \"Date\"\ncol2 (text): \"Opponent\"\ncol3 (text): \"Venue\"\ncol4 (text): \"Result\"\ncol5 (text): \"Attendance\"\n\nQuestion: What is the lowest game number on 20 July 2008?"}], "gold_sql": "SELECT MIN(col0) FROM data WHERE col1 = '20 july 2008';", "gold_answer": [1.0], "tokens": 205, "prompt_tokens": 181, "raw_output": "SELECT MIN(col0) FROM data WHERE col1 = '20 July 2008'", "seconds": 0.42494679195806384, "generation_tokens": 22, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [null], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5884, "table_id": "2-11786147-5", "question": "Who was the Home Team while Calgary was visiting while having an Attendance above 15,655?", "headers": ["Date", "Visitor", "Score", "Home", "Decision", "Attendance", "Record"], "types": ["text", "text", "text", "text", "text", "real", "text"], "query": {"sel": 3, "conds": [[1, 0, "calgary"], [5, 1, "15,655"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Visitor\"\ncol2 (text): \"Score\"\ncol3 (text): \"Home\"\ncol4 (text): \"Decision\"\ncol5 (real): \"Attendance\"\ncol6 (text): \"Record\"\n\nQuestion: Who was the Home Team while Calgary was visiting while having an Attendance above 15,655?"}], "gold_sql": "SELECT col3 FROM data WHERE col1 = 'calgary' AND col5 > 15655.0;", "gold_answer": ["colorado", "colorado"], "tokens": 221, "prompt_tokens": 193, "raw_output": "SELECT Home FROM data WHERE Decision = 'Home Team' AND Attendance > 15655 AND Home = 'Calgary'", "seconds": 0.20580524997785687, "generation_tokens": 28, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 8165, "table_id": "2-12410929-35", "question": "Which Home Captain has Eng by 208 runs?", "headers": ["Date", "Home captain", "Away captain", "Venue", "Result"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[4, 0, "eng by 208 runs"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Home captain\"\ncol2 (text): \"Away captain\"\ncol3 (text): \"Venue\"\ncol4 (text): \"Result\"\n\nQuestion: Which Home Captain has Eng by 208 runs?"}], "gold_sql": "SELECT col1 FROM data WHERE col4 = 'eng by 208 runs';", "gold_answer": ["richie richardson"], "tokens": 189, "prompt_tokens": 169, "raw_output": "SELECT col1 FROM data WHERE col2 = \"Eng\" AND col3 LIKE \"%208%\"", "seconds": 0.19501187512651086, "generation_tokens": 22, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 150, "table_id": "1-11734041-2", "question": "What school did the forward whose number is 10 belong to?", "headers": ["Player", "No.(s)", "Height in Ft.", "Position", "Years for Rockets", "School/Club Team/Country"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 5, "conds": [[1, 0, "10"], [3, 0, "Forward"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Player\"\ncol1 (text): \"No.(s)\"\ncol2 (text): \"Height in Ft.\"\ncol3 (text): \"Position\"\ncol4 (text): \"Years for Rockets\"\ncol5 (text): \"School/Club Team/Country\"\n\nQuestion: What school did the forward whose number is 10 belong to?"}], "gold_sql": "SELECT col5 FROM data WHERE col1 = '10' AND col3 = 'forward';", "gold_answer": ["arizona"], "tokens": 210, "prompt_tokens": 188, "raw_output": "SELECT T2.school FROM data AS T1 INNER JOIN school AS T2 ON T1.school = T2.school WHERE T1.no = 10", "seconds": 0.23960787476971745, "generation_tokens": 32, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5475, "table_id": "2-11750790-1", "question": "what is the sum of the losses when the goals against is less than 34 and the games played is less than 8?", "headers": ["Team", "Games Played", "Wins", "Losses", "Ties", "Goals For", "Goals Against"], "types": ["text", "real", "real", "real", "real", "real", "real"], "query": {"sel": 3, "conds": [[6, 2, 34], [1, 2, 8]], "agg": 4}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Team\"\ncol1 (real): \"Games Played\"\ncol2 (real): \"Wins\"\ncol3 (real): \"Losses\"\ncol4 (real): \"Ties\"\ncol5 (real): \"Goals For\"\ncol6 (real): \"Goals Against\"\n\nQuestion: what is the sum of the losses when the goals against is less than 34 and the games played is less than 8?"}], "gold_sql": "SELECT SUM(col3) FROM data WHERE col6 < 34 AND col1 < 8;", "gold_answer": [null], "tokens": 226, "prompt_tokens": 203, "raw_output": "SELECT SUM(losses) FROM data WHERE goals_against < 34 AND games_played < 8", "seconds": 0.2135888747870922, "generation_tokens": 23, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 755, "table_id": "1-16799784-7", "question": "what is the latitude of the feature of longitude 80.0e", "headers": ["Name", "Latitude", "Longitude", "Diameter (km)", "Year named", "Name origin"], "types": ["text", "text", "text", "text", "real", "text"], "query": {"sel": 1, "conds": [[2, 0, "80.0E"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"Latitude\"\ncol2 (text): \"Longitude\"\ncol3 (text): \"Diameter (km)\"\ncol4 (real): \"Year named\"\ncol5 (text): \"Name origin\"\n\nQuestion: what is the latitude of the feature of longitude 80.0e"}], "gold_sql": "SELECT col1 FROM data WHERE col2 = '80.0e';", "gold_answer": ["23.0s"], "tokens": 201, "prompt_tokens": 183, "raw_output": "SELECT col1 FROM data WHERE col2 = '80.0e'", "seconds": 0.17532570892944932, "generation_tokens": 17, "valid_sql": true, "answer_correct": true, "exact_sql": true, "predicted_answer": ["23.0s"], "answer_correct_without_fence": true}
{"source_split": "dev", "source_row": 1577, "table_id": "1-24222929-3", "question": "When did the episode that had 3.69 million total viewers (Live and SD types combined) first air?", "headers": ["Episode number Production number", "Title", "Original airdate", "Rating/Share in Households", "18–49 Rating Live+SD / Live+7", "Share", "Live+SD Total viewers", "Live+7 Day DVR Total viewers"], "types": ["text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[6, 0, "3.69 million"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Episode number Production number\"\ncol1 (text): \"Title\"\ncol2 (text): \"Original airdate\"\ncol3 (text): \"Rating/Share in Households\"\ncol4 (text): \"18–49 Rating Live+SD / Live+7\"\ncol5 (text): \"Share\"\ncol6 (text): \"Live+SD Total viewers\"\ncol7 (text): \"Live+7 Day DVR Total viewers\"\n\nQuestion: When did the episode that had 3.69 million total viewers (Live and SD types combined) first air?"}], "gold_sql": "SELECT col2 FROM data WHERE col6 = '3.69 million';", "gold_answer": ["december 16, 2009"], "tokens": 253, "prompt_tokens": 235, "raw_output": "SELECT col3 FROM data WHERE col6 = 'Live+SD Total viewers' AND col7 = '3.69 million' ORDER BY col0 ASC LIMIT 1", "seconds": 0.265191915910691, "generation_tokens": 36, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5936, "table_id": "2-11677760-25", "question": "How tall is Daniel Orton?", "headers": ["Player", "Height", "School", "Hometown", "College", "NBA Draft"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[0, 0, "daniel orton"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Player\"\ncol1 (text): \"Height\"\ncol2 (text): \"School\"\ncol3 (text): \"Hometown\"\ncol4 (text): \"College\"\ncol5 (text): \"NBA Draft\"\n\nQuestion: How tall is Daniel Orton?"}], "gold_sql": "SELECT col1 FROM data WHERE col0 = 'daniel orton';", "gold_answer": ["6-10"], "tokens": 190, "prompt_tokens": 173, "raw_output": "SELECT col1 FROM data WHERE col0 = 'Daniel Orton'", "seconds": 0.17251174990087748, "generation_tokens": 15, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1487, "table_id": "1-23465011-5", "question": "Which couple has 295 as a total score?", "headers": ["Average Ranking", "Competitive Finish", "Couple", "Number Of Dances", "Total Score", "Average"], "types": ["real", "real", "text", "real", "real", "text"], "query": {"sel": 2, "conds": [[4, 0, 295]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Average Ranking\"\ncol1 (real): \"Competitive Finish\"\ncol2 (text): \"Couple\"\ncol3 (real): \"Number Of Dances\"\ncol4 (real): \"Total Score\"\ncol5 (text): \"Average\"\n\nQuestion: Which couple has 295 as a total score?"}], "gold_sql": "SELECT col2 FROM data WHERE col4 = 295;", "gold_answer": ["patti & sandro"], "tokens": 199, "prompt_tokens": 183, "raw_output": "SELECT couple FROM data WHERE total_score = 295", "seconds": 0.1552434996701777, "generation_tokens": 13, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2137, "table_id": "1-27756314-11", "question": "How many different results for high rebounds were there for game number 76?", "headers": ["Game", "Date", "Team", "Score", "High points", "High rebounds", "High assists", "Location Attendance", "Record"], "types": ["real", "text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 5, "conds": [[0, 0, 76]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Game\"\ncol1 (text): \"Date\"\ncol2 (text): \"Team\"\ncol3 (text): \"Score\"\ncol4 (text): \"High points\"\ncol5 (text): \"High rebounds\"\ncol6 (text): \"High assists\"\ncol7 (text): \"Location Attendance\"\ncol8 (text): \"Record\"\n\nQuestion: How many different results for high rebounds were there for game number 76?"}], "gold_sql": "SELECT COUNT(col5) FROM data WHERE col0 = 76;", "gold_answer": [1], "tokens": 224, "prompt_tokens": 207, "raw_output": "SELECT COUNT(DISTINCT col6) FROM data WHERE col0 = 76 AND col1 = 'Game'", "seconds": 0.2063340418972075, "generation_tokens": 25, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7766, "table_id": "2-12865378-1", "question": "With an issue date(s) of 12 September, what is in the column for Weeks on Top?", "headers": ["Volume:Issue", "Issue Date(s)", "Weeks on Top", "Song", "Artist"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[1, 0, "12 september"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Volume:Issue\"\ncol1 (text): \"Issue Date(s)\"\ncol2 (text): \"Weeks on Top\"\ncol3 (text): \"Song\"\ncol4 (text): \"Artist\"\n\nQuestion: With an issue date(s) of 12 September, what is in the column for Weeks on Top?"}], "gold_sql": "SELECT col2 FROM data WHERE col1 = '12 september';", "gold_answer": ["1"], "tokens": 200, "prompt_tokens": 184, "raw_output": "SELECT col3 FROM data WHERE col1 = '12 September'", "seconds": 0.16738216718658805, "generation_tokens": 15, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4424, "table_id": "2-17430092-7", "question": "Which party does Timothy Kirkhope lead?", "headers": ["Leader", "Party", "Leader's Seat", "Number of Seats", "Leader since"], "types": ["text", "text", "text", "real", "text"], "query": {"sel": 1, "conds": [[0, 0, "timothy kirkhope"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Leader\"\ncol1 (text): \"Party\"\ncol2 (text): \"Leader's Seat\"\ncol3 (real): \"Number of Seats\"\ncol4 (text): \"Leader since\"\n\nQuestion: Which party does Timothy Kirkhope lead?"}], "gold_sql": "SELECT col1 FROM data WHERE col0 = 'timothy kirkhope';", "gold_answer": ["conservative party"], "tokens": 186, "prompt_tokens": 168, "raw_output": "SELECT Party FROM data WHERE Leader = 'Timothy Kirkhope'", "seconds": 0.153093374799937, "generation_tokens": 14, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1287, "table_id": "1-2182654-6", "question": "What was the name of the episode that was directed by Mary Harron?", "headers": ["No. in series", "No. in season", "Title", "Directed by", "Written by", "Original air date"], "types": ["real", "real", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[3, 0, "Mary Harron"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"No. in series\"\ncol1 (real): \"No. in season\"\ncol2 (text): \"Title\"\ncol3 (text): \"Directed by\"\ncol4 (text): \"Written by\"\ncol5 (text): \"Original air date\"\n\nQuestion: What was the name of the episode that was directed by Mary Harron?"}], "gold_sql": "SELECT col2 FROM data WHERE col3 = 'mary harron';", "gold_answer": ["\"the rainbow of her reasons\""], "tokens": 205, "prompt_tokens": 188, "raw_output": "SELECT T2.title FROM data AS T1 INNER JOIN director AS T2 ON T1.director = T2.name WHERE T1.director = 'Mary Harron'", "seconds": 0.2518065827898681, "generation_tokens": 36, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6282, "table_id": "2-1105741-2", "question": "What is the earliest year that had under 26 points and a toyota v8 engine?", "headers": ["Year", "Team", "Chassis", "Engine", "Tyres", "Points"], "types": ["real", "text", "text", "text", "text", "real"], "query": {"sel": 0, "conds": [[3, 0, "toyota v8"], [5, 2, 26]], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Year\"\ncol1 (text): \"Team\"\ncol2 (text): \"Chassis\"\ncol3 (text): \"Engine\"\ncol4 (text): \"Tyres\"\ncol5 (real): \"Points\"\n\nQuestion: What is the earliest year that had under 26 points and a toyota v8 engine?"}], "gold_sql": "SELECT MIN(col0) FROM data WHERE col3 = 'toyota v8' AND col5 < 26;", "gold_answer": [2006.0], "tokens": 212, "prompt_tokens": 185, "raw_output": "SELECT col0 FROM data WHERE col5 < 26 AND col2 = 'Toyota' AND col3 = 'V8' ORDER BY col0 ASC LIMIT 1", "seconds": 0.257939541246742, "generation_tokens": 36, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5736, "table_id": "2-11821711-9", "question": "What is the Decision listed when the Home was Colorado?", "headers": ["Date", "Visitor", "Score", "Home", "Decision", "Attendance", "Record"], "types": ["text", "text", "text", "text", "text", "real", "text"], "query": {"sel": 4, "conds": [[3, 0, "colorado"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Visitor\"\ncol2 (text): \"Score\"\ncol3 (text): \"Home\"\ncol4 (text): \"Decision\"\ncol5 (real): \"Attendance\"\ncol6 (text): \"Record\"\n\nQuestion: What is the Decision listed when the Home was Colorado?"}], "gold_sql": "SELECT col4 FROM data WHERE col3 = 'colorado';", "gold_answer": ["cloutier"], "tokens": 197, "prompt_tokens": 182, "raw_output": "SELECT Decision FROM data WHERE Home = \"Colorado\"", "seconds": 0.15506754210218787, "generation_tokens": 11, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3692, "table_id": "2-15759302-1", "question": "What is the Overall with a Date that is february 4, 1964?", "headers": ["Date", "Opponent", "Location", "Result", "Overall"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[0, 0, "february 4, 1964"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Opponent\"\ncol2 (text): \"Location\"\ncol3 (text): \"Result\"\ncol4 (text): \"Overall\"\n\nQuestion: What is the Overall with a Date that is february 4, 1964?"}], "gold_sql": "SELECT col4 FROM data WHERE col0 = 'february 4, 1964';", "gold_answer": ["6–6"], "tokens": 199, "prompt_tokens": 176, "raw_output": "SELECT Overall FROM data WHERE Date = '2019-02-04'", "seconds": 0.18759474996477365, "generation_tokens": 20, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5794, "table_id": "2-11775918-15", "question": "What college or league did the round 2 pick with d position come from?", "headers": ["Round", "Player", "Position", "Nationality", "College/Junior/Club Team (League)"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[0, 0, 2], [2, 0, "d"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Round\"\ncol1 (text): \"Player\"\ncol2 (text): \"Position\"\ncol3 (text): \"Nationality\"\ncol4 (text): \"College/Junior/Club Team (League)\"\n\nQuestion: What college or league did the round 2 pick with d position come from?"}], "gold_sql": "SELECT col4 FROM data WHERE col0 = 2 AND col2 = 'd';", "gold_answer": ["gaylord high school ( ushs-mi )"], "tokens": 200, "prompt_tokens": 180, "raw_output": "SELECT college FROM data WHERE round = 2 AND position = 'd'", "seconds": 0.1645754580385983, "generation_tokens": 16, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1180, "table_id": "1-20925393-3", "question": "state the name of day in english where cardinal direction is east", "headers": ["Cardinal direction", "Burmese", "Sanskrit", "English", "Planet", "Sign"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[0, 0, "East"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Cardinal direction\"\ncol1 (text): \"Burmese\"\ncol2 (text): \"Sanskrit\"\ncol3 (text): \"English\"\ncol4 (text): \"Planet\"\ncol5 (text): \"Sign\"\n\nQuestion: state the name of day in english where cardinal direction is east"}], "gold_sql": "SELECT col3 FROM data WHERE col0 = 'east';", "gold_answer": ["monday"], "tokens": 195, "prompt_tokens": 181, "raw_output": "SELECT English FROM data WHERE Cardinal direction = \"East\"", "seconds": 0.14831045782193542, "generation_tokens": 12, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Invalid expression / Unexpected token. Line 1, Col: 49.\n  SELECT English FROM data WHERE Cardinal \u001b[4mdirection\u001b[0m = \"East\"", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1469, "table_id": "1-23286223-8", "question": "Where was the March 24 game played?", "headers": ["Game", "Date", "Team", "Score", "High points", "High rebounds", "High assists", "Location Attendance", "Record"], "types": ["real", "text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 7, "conds": [[1, 0, "March 24"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Game\"\ncol1 (text): \"Date\"\ncol2 (text): \"Team\"\ncol3 (text): \"Score\"\ncol4 (text): \"High points\"\ncol5 (text): \"High rebounds\"\ncol6 (text): \"High assists\"\ncol7 (text): \"Location Attendance\"\ncol8 (text): \"Record\"\n\nQuestion: Where was the March 24 game played?"}], "gold_sql": "SELECT col7 FROM data WHERE col1 = 'march 24';", "gold_answer": ["air canada centre 16,178"], "tokens": 219, "prompt_tokens": 201, "raw_output": "SELECT col2 FROM data WHERE col1 = 'March 24'", "seconds": 0.16964675020426512, "generation_tokens": 16, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 462, "table_id": "1-14342480-6", "question": "Name the fewest touchdowns", "headers": ["Player", "Position", "Starter", "Touchdowns", "Extra points", "Field goals", "Points"], "types": ["text", "text", "text", "real", "real", "real", "real"], "query": {"sel": 3, "conds": [], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Player\"\ncol1 (text): \"Position\"\ncol2 (text): \"Starter\"\ncol3 (real): \"Touchdowns\"\ncol4 (real): \"Extra points\"\ncol5 (real): \"Field goals\"\ncol6 (real): \"Points\"\n\nQuestion: Name the fewest touchdowns"}], "gold_sql": "SELECT MIN(col3) FROM data;", "gold_answer": [1.0], "tokens": 191, "prompt_tokens": 181, "raw_output": "SELECT col0 FROM data ORDER BY col3 LIMIT 1", "seconds": 0.1493219998665154, "generation_tokens": 13, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5674, "table_id": "2-11570261-11", "question": "Who is Runner(s)-up that has a Date of may 24, 1999?", "headers": ["Date", "Tournament", "Winning score", "Margin of victory", "Runner(s)-up"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[0, 0, "may 24, 1999"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Tournament\"\ncol2 (text): \"Winning score\"\ncol3 (text): \"Margin of victory\"\ncol4 (text): \"Runner(s)-up\"\n\nQuestion: Who is Runner(s)-up that has a Date of may 24, 1999?"}], "gold_sql": "SELECT col4 FROM data WHERE col0 = 'may 24, 1999';", "gold_answer": ["retief goosen"], "tokens": 208, "prompt_tokens": 185, "raw_output": "SELECT Runner(s)-up FROM data WHERE Date = '1999-05-24'", "seconds": 0.1927262912504375, "generation_tokens": 23, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5492, "table_id": "2-11590808-1", "question": "What is the low code for glen harvie with a population greater than 2,774?", "headers": ["Place", "Code", "Area (km 2 )", "Population", "Most spoken language"], "types": ["text", "real", "real", "real", "text"], "query": {"sel": 1, "conds": [[3, 1, "2,774"], [0, 0, "glen harvie"]], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Place\"\ncol1 (real): \"Code\"\ncol2 (real): \"Area (km 2 )\"\ncol3 (real): \"Population\"\ncol4 (text): \"Most spoken language\"\n\nQuestion: What is the low code for glen harvie with a population greater than 2,774?"}], "gold_sql": "SELECT MIN(col1) FROM data WHERE col3 > 2774.0 AND col0 = 'glen harvie';", "gold_answer": [70305.0], "tokens": 214, "prompt_tokens": 184, "raw_output": "SELECT col1 FROM data WHERE col0 = 'Glen Harvie' AND col2 > 2774", "seconds": 0.2093342086300254, "generation_tokens": 25, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2055, "table_id": "1-27495117-3", "question": "how many teams had an appointment date of 11 april 2011", "headers": ["Team", "Outgoing manager", "Manner of departure", "Date of vacancy", "Replaced by", "Date of appointment", "Position in table"], "types": ["text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[5, 0, "11 April 2011"]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Team\"\ncol1 (text): \"Outgoing manager\"\ncol2 (text): \"Manner of departure\"\ncol3 (text): \"Date of vacancy\"\ncol4 (text): \"Replaced by\"\ncol5 (text): \"Date of appointment\"\ncol6 (text): \"Position in table\"\n\nQuestion: how many teams had an appointment date of 11 april 2011"}], "gold_sql": "SELECT COUNT(col0) FROM data WHERE col5 = '11 april 2011';", "gold_answer": [1], "tokens": 224, "prompt_tokens": 201, "raw_output": "SELECT COUNT(*) FROM data WHERE date_of_appointment = '11 april 2011'", "seconds": 0.20386174973100424, "generation_tokens": 22, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6389, "table_id": "2-1123405-2", "question": "How many laps did Ricardo Zonta have?", "headers": ["Driver", "Constructor", "Laps", "Time/Retired", "Grid"], "types": ["text", "text", "real", "text", "real"], "query": {"sel": 2, "conds": [[0, 0, "ricardo zonta"]], "agg": 4}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Driver\"\ncol1 (text): \"Constructor\"\ncol2 (real): \"Laps\"\ncol3 (text): \"Time/Retired\"\ncol4 (real): \"Grid\"\n\nQuestion: How many laps did Ricardo Zonta have?"}], "gold_sql": "SELECT SUM(col2) FROM data WHERE col0 = 'ricardo zonta';", "gold_answer": [53.0], "tokens": 187, "prompt_tokens": 168, "raw_output": "SELECT COUNT(*) FROM data WHERE col0 = 'Ricardo Zonta'", "seconds": 0.16423604218289256, "generation_tokens": 17, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [0], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6861, "table_id": "2-17759945-8", "question": "What is the position of the player with a height of m (ft 6in)?", "headers": ["Name", "Pos.", "Height", "Weight", "Date of Birth", "Club"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[2, 0, "m (ft 6in)"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"Pos.\"\ncol2 (text): \"Height\"\ncol3 (text): \"Weight\"\ncol4 (text): \"Date of Birth\"\ncol5 (text): \"Club\"\n\nQuestion: What is the position of the player with a height of m (ft 6in)?"}], "gold_sql": "SELECT col1 FROM data WHERE col2 = 'm (ft 6in)';", "gold_answer": ["d"], "tokens": 203, "prompt_tokens": 183, "raw_output": "SELECT col1 FROM data WHERE col2 LIKE 'm%'", "seconds": 0.15541099989786744, "generation_tokens": 13, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3666, "table_id": "2-15442855-1", "question": "Which Label has a Date smaller than 2008, and a Catalog of fall cd 006?", "headers": ["Region", "Date", "Label", "Format", "Catalog"], "types": ["text", "real", "text", "text", "text"], "query": {"sel": 2, "conds": [[1, 2, 2008], [4, 0, "fall cd 006"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Region\"\ncol1 (real): \"Date\"\ncol2 (text): \"Label\"\ncol3 (text): \"Format\"\ncol4 (text): \"Catalog\"\n\nQuestion: Which Label has a Date smaller than 2008, and a Catalog of fall cd 006?"}], "gold_sql": "SELECT col2 FROM data WHERE col1 < 2008 AND col4 = 'fall cd 006';", "gold_answer": ["fall out records"], "tokens": 207, "prompt_tokens": 179, "raw_output": "SELECT Label FROM data WHERE Date < 2008 AND Catalog = 'fall cd 006'", "seconds": 0.19789624959230423, "generation_tokens": 24, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1560, "table_id": "1-24172157-3", "question": "What is the date of vacancy for the Liverpool team with a table named pre-season?", "headers": ["Team", "Outgoing manager", "Manner of departure", "Date of vacancy", "Table", "Incoming manager", "Date of appointment"], "types": ["text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[4, 0, "Pre-season"], [0, 0, "Liverpool"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Team\"\ncol1 (text): \"Outgoing manager\"\ncol2 (text): \"Manner of departure\"\ncol3 (text): \"Date of vacancy\"\ncol4 (text): \"Table\"\ncol5 (text): \"Incoming manager\"\ncol6 (text): \"Date of appointment\"\n\nQuestion: What is the date of vacancy for the Liverpool team with a table named pre-season?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 = 'pre-season' AND col0 = 'liverpool';", "gold_answer": ["3 june 2010"], "tokens": 222, "prompt_tokens": 198, "raw_output": "SELECT Date_of_vacancy FROM data WHERE Team = 'Liverpool' AND Table = 'pre-season'", "seconds": 0.19529949966818094, "generation_tokens": 22, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3641, "table_id": "2-15988551-1", "question": "What is the number of totals that have silvers under 2, bronzes over 0, and golds over 1?", "headers": ["Rank", "Nation", "Gold", "Silver", "Bronze", "Total"], "types": ["text", "text", "real", "real", "real", "real"], "query": {"sel": 5, "conds": [[3, 2, 2], [4, 1, 0], [2, 1, 1]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Rank\"\ncol1 (text): \"Nation\"\ncol2 (real): \"Gold\"\ncol3 (real): \"Silver\"\ncol4 (real): \"Bronze\"\ncol5 (real): \"Total\"\n\nQuestion: What is the number of totals that have silvers under 2, bronzes over 0, and golds over 1?"}], "gold_sql": "SELECT COUNT(col5) FROM data WHERE col3 < 2 AND col4 > 0 AND col2 > 1;", "gold_answer": [0], "tokens": 220, "prompt_tokens": 192, "raw_output": "SELECT COUNT(T1.col5) FROM data AS T1 INNER JOIN data AS T2 ON T1.col0 = T2.col0 WHERE T2.col1 = 'Nation' AND T2.col2 < 2 AND T2.col3 > 0 AND T2.col4 > 1", "seconds": 0.37026720913127065, "generation_tokens": 62, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4741, "table_id": "2-16575609-4", "question": "Which College has a Position of ol, and a Pick # smaller than 32?", "headers": ["Pick #", "CFL Team", "Player", "Position", "College"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[3, 0, "ol"], [0, 2, 32]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Pick #\"\ncol1 (text): \"CFL Team\"\ncol2 (text): \"Player\"\ncol3 (text): \"Position\"\ncol4 (text): \"College\"\n\nQuestion: Which College has a Position of ol, and a Pick # smaller than 32?"}], "gold_sql": "SELECT col4 FROM data WHERE col3 = 'ol' AND col0 < 32;", "gold_answer": ["montréal"], "tokens": 198, "prompt_tokens": 176, "raw_output": "SELECT College FROM data WHERE Position = 'ol' AND Pick # < 32", "seconds": 0.1874122079461813, "generation_tokens": 18, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Invalid expression / Unexpected token. Line 1, Col: 57.\n  SELECT College FROM data WHERE Position = 'ol' AND Pick \u001b[4m#\u001b[0m < 32", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2000, "table_id": "1-26998693-1", "question": "How much were the f/laps if poles is higher than 1.0 during 2008?", "headers": ["Season", "Series", "Team", "Races", "Wins", "Poles", "F/Laps", "Podiums", "Points", "Position"], "types": ["real", "text", "text", "real", "real", "real", "real", "real", "text", "text"], "query": {"sel": 6, "conds": [[5, 1, 1.0], [0, 0, 2008]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Season\"\ncol1 (text): \"Series\"\ncol2 (text): \"Team\"\ncol3 (real): \"Races\"\ncol4 (real): \"Wins\"\ncol5 (real): \"Poles\"\ncol6 (real): \"F/Laps\"\ncol7 (real): \"Podiums\"\ncol8 (text): \"Points\"\ncol9 (text): \"Position\"\n\nQuestion: How much were the f/laps if poles is higher than 1.0 during 2008?"}], "gold_sql": "SELECT col6 FROM data WHERE col5 > 1.0 AND col0 = 2008;", "gold_answer": [1.0], "tokens": 249, "prompt_tokens": 224, "raw_output": "SELECT SUM(T2.F/Laps) FROM data AS T1 INNER JOIN data AS T2 ON T1.col0 = T2.col0 WHERE T1.col1 = 'Series' AND T1.col2 = 'Team' AND T1.col3 = 2008 AND T1.col4 > 1.0 AND T2.col6 = 'F/Laps'", "seconds": 0.4493484999984503, "generation_tokens": 80, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2800, "table_id": "2-1463479-1", "question": "which open cup was in 2012?", "headers": ["Year", "Division", "League", "Reg. Season", "Playoffs", "Open Cup"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 5, "conds": [[0, 0, "2012"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Year\"\ncol1 (text): \"Division\"\ncol2 (text): \"League\"\ncol3 (text): \"Reg. Season\"\ncol4 (text): \"Playoffs\"\ncol5 (text): \"Open Cup\"\n\nQuestion: which open cup was in 2012?"}], "gold_sql": "SELECT col5 FROM data WHERE col0 = '2012';", "gold_answer": ["did not qualify"], "tokens": 195, "prompt_tokens": 178, "raw_output": "SELECT col5 FROM data WHERE col4 = 'Open Cup' AND YEAR = 2012", "seconds": 0.20614904211834073, "generation_tokens": 22, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7169, "table_id": "2-18662685-12", "question": "What country has a rank smaller than 6, a time of 6:32.32 and notes of FB?", "headers": ["Rank", "Rowers", "Country", "Time", "Notes"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[0, 2, 6], [4, 0, "fb"], [3, 0, "6:32.32"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Rank\"\ncol1 (text): \"Rowers\"\ncol2 (text): \"Country\"\ncol3 (text): \"Time\"\ncol4 (text): \"Notes\"\n\nQuestion: What country has a rank smaller than 6, a time of 6:32.32 and notes of FB?"}], "gold_sql": "SELECT col2 FROM data WHERE col0 < 6 AND col4 = 'fb' AND col3 = '6:32.32';", "gold_answer": ["australia"], "tokens": 215, "prompt_tokens": 182, "raw_output": "SELECT country FROM data WHERE rank < 6 AND time = '6:32.32' AND notes = 'FB'", "seconds": 0.23306729178875685, "generation_tokens": 28, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3197, "table_id": "2-15229416-3", "question": "What is the accession number of the protein with a divergence from human lineage of 937.5?", "headers": ["common name", "protein name", "divergence from human lineage (MYA)", "accession number", "sequence length (aa)", "sequence identity to human protein"], "types": ["text", "text", "real", "text", "real", "text"], "query": {"sel": 3, "conds": [[2, 0, 937.5]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"common name\"\ncol1 (text): \"protein name\"\ncol2 (real): \"divergence from human lineage (MYA)\"\ncol3 (text): \"accession number\"\ncol4 (real): \"sequence length (aa)\"\ncol5 (text): \"sequence identity to human protein\"\n\nQuestion: What is the accession number of the protein with a divergence from human lineage of 937.5?"}], "gold_sql": "SELECT col3 FROM data WHERE col2 = 937.5;", "gold_answer": ["xp_003380263.1"], "tokens": 222, "prompt_tokens": 204, "raw_output": "SELECT accession_number FROM data WHERE divergence_from_human_lineage = 937.5", "seconds": 0.18772195791825652, "generation_tokens": 19, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4323, "table_id": "2-17029703-3", "question": "What is the average number of bronze of the nation with more than 1 gold and 1 silver medal?", "headers": ["Rank", "Nation", "Gold", "Silver", "Bronze", "Total"], "types": ["text", "text", "real", "real", "real", "real"], "query": {"sel": 4, "conds": [[2, 1, 1], [3, 0, 1]], "agg": 5}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Rank\"\ncol1 (text): \"Nation\"\ncol2 (real): \"Gold\"\ncol3 (real): \"Silver\"\ncol4 (real): \"Bronze\"\ncol5 (real): \"Total\"\n\nQuestion: What is the average number of bronze of the nation with more than 1 gold and 1 silver medal?"}], "gold_sql": "SELECT AVG(col4) FROM data WHERE col2 > 1 AND col3 = 1;", "gold_answer": [2.0], "tokens": 208, "prompt_tokens": 186, "raw_output": "SELECT AVG(T2.Bronze) FROM data AS T1 INNER JOIN data AS T2 ON T1.col0 = T2.col0 WHERE T1.col1 = 'United States' AND T1.col2 > 1 AND T1.col3 > 1", "seconds": 0.33207145892083645, "generation_tokens": 56, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1452, "table_id": "1-23285805-4", "question": "If the record is 5-5, what is the game maximum?", "headers": ["Game", "Date", "Team", "Score", "High points", "High rebounds", "High assists", "Location Attendance", "Record"], "types": ["real", "text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[8, 0, "5-5"]], "agg": 1}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Game\"\ncol1 (text): \"Date\"\ncol2 (text): \"Team\"\ncol3 (text): \"Score\"\ncol4 (text): \"High points\"\ncol5 (text): \"High rebounds\"\ncol6 (text): \"High assists\"\ncol7 (text): \"Location Attendance\"\ncol8 (text): \"Record\"\n\nQuestion: If the record is 5-5, what is the game maximum?"}], "gold_sql": "SELECT MAX(col0) FROM data WHERE col8 = '5-5';", "gold_answer": [10.0], "tokens": 224, "prompt_tokens": 206, "raw_output": "SELECT MAX(col0) FROM data WHERE col7 = '5-5'", "seconds": 0.1890569580718875, "generation_tokens": 17, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [null], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 8128, "table_id": "2-12264570-2", "question": "When doeas Mascot of blue devils in Gary Froebel School?", "headers": ["School", "Location", "Mascot", "County", "Joined", "Previous Conference", "Left", "Conference joined"], "types": ["text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[2, 0, "blue devils"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"School\"\ncol1 (text): \"Location\"\ncol2 (text): \"Mascot\"\ncol3 (text): \"County\"\ncol4 (text): \"Joined\"\ncol5 (text): \"Previous Conference\"\ncol6 (text): \"Left\"\ncol7 (text): \"Conference joined\"\n\nQuestion: When doeas Mascot of blue devils in Gary Froebel School?"}], "gold_sql": "SELECT col4 FROM data WHERE col2 = 'blue devils';", "gold_answer": ["1927"], "tokens": 215, "prompt_tokens": 199, "raw_output": "SELECT MAX(Left) FROM data WHERE School = 'Gary Froebel School' AND Mascot = 'Blue Devils'", "seconds": 0.22461525024846196, "generation_tokens": 27, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3688, "table_id": "2-15344743-1", "question": "Which builder completed before 1890 and launched on 9 june 1888?", "headers": ["Name", "Builder", "Laid down", "Launched", "Completed"], "types": ["text", "text", "text", "text", "real"], "query": {"sel": 1, "conds": [[4, 2, 1890], [3, 0, "9 june 1888"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"Builder\"\ncol2 (text): \"Laid down\"\ncol3 (text): \"Launched\"\ncol4 (real): \"Completed\"\n\nQuestion: Which builder completed before 1890 and launched on 9 june 1888?"}], "gold_sql": "SELECT col1 FROM data WHERE col4 < 1890 AND col3 = '9 june 1888';", "gold_answer": ["chatham"], "tokens": 210, "prompt_tokens": 180, "raw_output": "SELECT Builder FROM data WHERE Builder = \"John Doe\" AND Launch > 1890 AND Launched = 9", "seconds": 0.21474237507209182, "generation_tokens": 26, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4250, "table_id": "2-1730774-6", "question": "What is Case Suffix (Case), when English Meaning is \"to Georgia, in Georgia\"?", "headers": ["noun root (meaning)", "case suffix (case)", "postposition", "full word", "English meaning"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[4, 0, "to georgia, in georgia"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"noun root (meaning)\"\ncol1 (text): \"case suffix (case)\"\ncol2 (text): \"postposition\"\ncol3 (text): \"full word\"\ncol4 (text): \"English meaning\"\n\nQuestion: What is Case Suffix (Case), when English Meaning is \"to Georgia, in Georgia\"?"}], "gold_sql": "SELECT col1 FROM data WHERE col4 = 'to georgia, in georgia';", "gold_answer": ["-s (dative)"], "tokens": 203, "prompt_tokens": 183, "raw_output": "SELECT col1 FROM data WHERE col2 = \"Case\" AND col3 = \"to Georgia, in Georgia\"", "seconds": 0.2077856669202447, "generation_tokens": 24, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4869, "table_id": "2-17167467-3", "question": "Which label has a year older than 2004 and a 2xcd detail as well as the sonic seducer cold hands seduction vol. 69 title?", "headers": ["Title", "Details", "Track Title", "Label", "Year"], "types": ["text", "text", "text", "text", "real"], "query": {"sel": 3, "conds": [[4, 1, 2004], [1, 0, "2xcd"], [0, 0, "sonic seducer cold hands seduction vol. 69"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Title\"\ncol1 (text): \"Details\"\ncol2 (text): \"Track Title\"\ncol3 (text): \"Label\"\ncol4 (real): \"Year\"\n\nQuestion: Which label has a year older than 2004 and a 2xcd detail as well as the sonic seducer cold hands seduction vol. 69 title?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 > 2004 AND col1 = '2xcd' AND col0 = 'sonic seducer cold hands seduction vol. 69';", "gold_answer": ["sonic seducer"], "tokens": 235, "prompt_tokens": 192, "raw_output": "SELECT label FROM data WHERE year > 2004 AND details LIKE '%2xcd%' AND title = 'sonic seducer cold hands seduction vol. 69'", "seconds": 0.2608677498064935, "generation_tokens": 38, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7636, "table_id": "2-1851542-3", "question": "What is the fewest wins for Thomas in events he had entered exactly 9 times?", "headers": ["Tournament", "Wins", "Top-5", "Top-25", "Events", "Cuts made"], "types": ["text", "real", "real", "real", "real", "real"], "query": {"sel": 1, "conds": [[4, 0, 9]], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Tournament\"\ncol1 (real): \"Wins\"\ncol2 (real): \"Top-5\"\ncol3 (real): \"Top-25\"\ncol4 (real): \"Events\"\ncol5 (real): \"Cuts made\"\n\nQuestion: What is the fewest wins for Thomas in events he had entered exactly 9 times?"}], "gold_sql": "SELECT MIN(col1) FROM data WHERE col4 = 9;", "gold_answer": [0.0], "tokens": 205, "prompt_tokens": 189, "raw_output": "SELECT col0 FROM data WHERE col1 = 9 AND col5 = \"Tournament\"", "seconds": 0.19461370911449194, "generation_tokens": 20, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 8378, "table_id": "2-12631319-1", "question": "How many carries have an average under 8.7 and touchdowns of 72?", "headers": ["Year", "Carries", "Rushing Yards", "Average", "Touchdowns", "Points"], "types": ["text", "real", "real", "real", "real", "real"], "query": {"sel": 1, "conds": [[3, 2, 8.7], [4, 0, 72]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Year\"\ncol1 (real): \"Carries\"\ncol2 (real): \"Rushing Yards\"\ncol3 (real): \"Average\"\ncol4 (real): \"Touchdowns\"\ncol5 (real): \"Points\"\n\nQuestion: How many carries have an average under 8.7 and touchdowns of 72?"}], "gold_sql": "SELECT COUNT(col1) FROM data WHERE col3 < 8.7 AND col4 = 72;", "gold_answer": [1], "tokens": 212, "prompt_tokens": 187, "raw_output": "SELECT COUNT(col0) FROM data WHERE col3 < 8.7 AND col4 = 72", "seconds": 0.2041734172962606, "generation_tokens": 23, "valid_sql": true, "answer_correct": true, "exact_sql": false, "predicted_answer": [1], "answer_correct_without_fence": true}
{"source_split": "dev", "source_row": 6548, "table_id": "2-17840654-3", "question": "In what category was Herself nominated?", "headers": ["Year", "Award", "Category", "Nominated", "Result"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[3, 0, "herself"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Year\"\ncol1 (text): \"Award\"\ncol2 (text): \"Category\"\ncol3 (text): \"Nominated\"\ncol4 (text): \"Result\"\n\nQuestion: In what category was Herself nominated?"}], "gold_sql": "SELECT col2 FROM data WHERE col3 = 'herself';", "gold_answer": ["twitter star of the year"], "tokens": 180, "prompt_tokens": 164, "raw_output": "SELECT col2 FROM data WHERE col0 = 'Herself'", "seconds": 0.16132841631770134, "generation_tokens": 15, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6547, "table_id": "2-17840654-3", "question": "Name the performance nominated for a Capif Award.", "headers": ["Year", "Award", "Category", "Nominated", "Result"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[1, 0, "capif award"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Year\"\ncol1 (text): \"Award\"\ncol2 (text): \"Category\"\ncol3 (text): \"Nominated\"\ncol4 (text): \"Result\"\n\nQuestion: Name the performance nominated for a Capif Award."}], "gold_sql": "SELECT col3 FROM data WHERE col1 = 'capif award';", "gold_answer": ["teen angels 2"], "tokens": 182, "prompt_tokens": 166, "raw_output": "SELECT col1 FROM data WHERE col2 = 'Capif'", "seconds": 0.15276404190808535, "generation_tokens": 14, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3587, "table_id": "2-1478239-7", "question": "What percentage of people were opposed to the candidate based on the Time Poll poll that showed 6% of people were unsure?", "headers": ["Poll Source", "Date", "Candidate", "Support", "Consider", "Oppose", "Unsure"], "types": ["text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 5, "conds": [[0, 0, "time poll"], [6, 0, "6%"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Poll Source\"\ncol1 (text): \"Date\"\ncol2 (text): \"Candidate\"\ncol3 (text): \"Support\"\ncol4 (text): \"Consider\"\ncol5 (text): \"Oppose\"\ncol6 (text): \"Unsure\"\n\nQuestion: What percentage of people were opposed to the candidate based on the Time Poll poll that showed 6% of people were unsure?"}], "gold_sql": "SELECT col5 FROM data WHERE col0 = 'time poll' AND col6 = '6%';", "gold_answer": ["34%"], "tokens": 222, "prompt_tokens": 199, "raw_output": "SELECT CAST(SUM(CASE WHEN col6 = 'Oppose' THEN 1 ELSE 0 END) AS REAL) * 100 / COUNT(*) FROM data WHERE col5 = 'Time Poll' AND col6 = 'Oppose' AND col4 = '6%'", "seconds": 0.358453958760947, "generation_tokens": 59, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 831, "table_id": "1-17356042-1", "question": "Name the title that was written by r. scott gemmill", "headers": ["Series #", "Season #", "Title", "Directed by", "Written by", "Original air date"], "types": ["real", "real", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[4, 0, "R. Scott Gemmill"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Series #\"\ncol1 (real): \"Season #\"\ncol2 (text): \"Title\"\ncol3 (text): \"Directed by\"\ncol4 (text): \"Written by\"\ncol5 (text): \"Original air date\"\n\nQuestion: Name the title that was written by r. scott gemmill"}], "gold_sql": "SELECT col2 FROM data WHERE col4 = 'r. scott gemmill';", "gold_answer": ["\"try carter\""], "tokens": 201, "prompt_tokens": 182, "raw_output": "SELECT title FROM data WHERE directed_by = 'R. Scott Gemmill'", "seconds": 0.1823822921141982, "generation_tokens": 16, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5847, "table_id": "2-1126516-1", "question": "When was Alianza's first season in first division with a promotion after 1959?", "headers": ["Name", "City", "Stadium", "First season in first division", "First season after most recent promotion"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[4, 0, "1959"], [0, 0, "alianza"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"City\"\ncol2 (text): \"Stadium\"\ncol3 (text): \"First season in first division\"\ncol4 (text): \"First season after most recent promotion\"\n\nQuestion: When was Alianza's first season in first division with a promotion after 1959?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 = '1959' AND col0 = 'alianza';", "gold_answer": ["1959"], "tokens": 210, "prompt_tokens": 185, "raw_output": "SELECT col4 FROM data WHERE col0 = 'Alianza' AND col1 = 'City' AND col2 = 'Stadium' AND col3 = 'First season in first division' AND col4 > '1959'", "seconds": 0.3190064998343587, "generation_tokens": 50, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 211, "table_id": "1-1211545-2", "question": "What is the toll for heavy vehicles with 3/4 axles at Verkeerdevlei toll plaza?", "headers": ["Name", "Location", "Light vehicle", "Heavy vehicle (2 axles)", "Heavy vehicle (3/4 axles)", "Heavy vehicle (5+ axles)"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[0, 0, "Verkeerdevlei Toll Plaza"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"Location\"\ncol2 (text): \"Light vehicle\"\ncol3 (text): \"Heavy vehicle (2 axles)\"\ncol4 (text): \"Heavy vehicle (3/4 axles)\"\ncol5 (text): \"Heavy vehicle (5+ axles)\"\n\nQuestion: What is the toll for heavy vehicles with 3/4 axles at Verkeerdevlei toll plaza?"}], "gold_sql": "SELECT col4 FROM data WHERE col0 = 'verkeerdevlei toll plaza';", "gold_answer": ["r117.00"], "tokens": 225, "prompt_tokens": 205, "raw_output": "SELECT col5 FROM data WHERE col2 = \"Heavy vehicle (3/4 axles)\" AND col3 = \"Verkeerdevlei toll plaza\"", "seconds": 0.25249520782381296, "generation_tokens": 33, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 8397, "table_id": "2-12886178-4", "question": "How many tries did the club with a try bonus of correct as of 2 June 2009 have?", "headers": ["Club", "Played", "Drawn", "Lost", "Points for", "Points against", "Tries for", "Tries against", "Try bonus", "Losing bonus", "Points"], "types": ["text", "text", "text", "text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 6, "conds": [[8, 0, "correct as of 2 june 2009"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Club\"\ncol1 (text): \"Played\"\ncol2 (text): \"Drawn\"\ncol3 (text): \"Lost\"\ncol4 (text): \"Points for\"\ncol5 (text): \"Points against\"\ncol6 (text): \"Tries for\"\ncol7 (text): \"Tries against\"\ncol8 (text): \"Try bonus\"\ncol9 (text): \"Losing bonus\"\ncol10 (text): \"Points\"\n\nQuestion: How many tries did the club with a try bonus of correct as of 2 June 2009 have?"}], "gold_sql": "SELECT col6 FROM data WHERE col8 = 'correct as of 2 june 2009';", "gold_answer": ["correct as of 2 june 2009"], "tokens": 263, "prompt_tokens": 238, "raw_output": "SELECT SUM(Tries_for) FROM data WHERE col9 = 'Correct' AND col10 = 'Try bonus' AND col2 = '2 June 2009'", "seconds": 0.2780337920412421, "generation_tokens": 38, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6705, "table_id": "2-18178534-6", "question": "What's the status on 16/12/1995?", "headers": ["Opposing Teams", "Against", "Date", "Venue", "Status"], "types": ["text", "real", "text", "text", "text"], "query": {"sel": 4, "conds": [[2, 0, "16/12/1995"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Opposing Teams\"\ncol1 (real): \"Against\"\ncol2 (text): \"Date\"\ncol3 (text): \"Venue\"\ncol4 (text): \"Status\"\n\nQuestion: What's the status on 16/12/1995?"}], "gold_sql": "SELECT col4 FROM data WHERE col2 = '16/12/1995';", "gold_answer": ["test match"], "tokens": 197, "prompt_tokens": 174, "raw_output": "SELECT Status FROM data WHERE Date = '16/12/1995'", "seconds": 0.18997745821252465, "generation_tokens": 20, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5453, "table_id": "2-10856203-2", "question": "What was the margin of victory when Brandt Snedeker was runner-up?", "headers": ["Date", "Tournament", "Winning score", "To par", "Margin of victory", "Runner-up"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[5, 0, "brandt snedeker"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Tournament\"\ncol2 (text): \"Winning score\"\ncol3 (text): \"To par\"\ncol4 (text): \"Margin of victory\"\ncol5 (text): \"Runner-up\"\n\nQuestion: What was the margin of victory when Brandt Snedeker was runner-up?"}], "gold_sql": "SELECT col4 FROM data WHERE col5 = 'brandt snedeker';", "gold_answer": ["3 strokes"], "tokens": 204, "prompt_tokens": 186, "raw_output": "SELECT Margin_of_victory FROM data WHERE Runner-up = 'Brandt Snedeker'", "seconds": 0.18862691707909107, "generation_tokens": 19, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1207, "table_id": "1-21350772-2", "question": "What is the 100 year when 500 year is 153?", "headers": ["Gas name", "Chemical formula", "Lifetime (years)", "20-yr", "100-yr", "500-yr"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[5, 0, "153"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Gas name\"\ncol1 (text): \"Chemical formula\"\ncol2 (text): \"Lifetime (years)\"\ncol3 (text): \"20-yr\"\ncol4 (text): \"100-yr\"\ncol5 (text): \"500-yr\"\n\nQuestion: What is the 100 year when 500 year is 153?"}], "gold_sql": "SELECT col4 FROM data WHERE col5 = '153';", "gold_answer": ["298"], "tokens": 215, "prompt_tokens": 199, "raw_output": "SELECT col4 FROM data WHERE col5 = '500-yr' AND col2 = '153' ORDER BY col4 LIMIT 1", "seconds": 0.24508141726255417, "generation_tokens": 33, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5267, "table_id": "2-171680-1", "question": "What rank has a gross of $35,976,000?", "headers": ["Rank", "Title", "Studio", "Director(s)", "Gross"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[4, 0, "$35,976,000"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Rank\"\ncol1 (text): \"Title\"\ncol2 (text): \"Studio\"\ncol3 (text): \"Director(s)\"\ncol4 (text): \"Gross\"\n\nQuestion: What rank has a gross of $35,976,000?"}], "gold_sql": "SELECT col0 FROM data WHERE col4 = '$35,976,000';", "gold_answer": [4.0], "tokens": 198, "prompt_tokens": 175, "raw_output": "SELECT col0 FROM data WHERE col3 = \"Studio\" AND col4 = \"Gross\" AND col1 = \"Title\" AND col0 = (SELECT MAX(col0) FROM data WHERE col3 = \"Studio\" AND col4 = \"Gross\" AND col1 = \"Title\")", "seconds": 0.37215625029057264, "generation_tokens": 62, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 516, "table_id": "1-14850099-18", "question": "What is the Burmese term associated with a cardinal direction of west?", "headers": ["Cardinal direction", "Burmese", "Sanskrit", "English", "Planet", "Sign"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[0, 0, "West"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Cardinal direction\"\ncol1 (text): \"Burmese\"\ncol2 (text): \"Sanskrit\"\ncol3 (text): \"English\"\ncol4 (text): \"Planet\"\ncol5 (text): \"Sign\"\n\nQuestion: What is the Burmese term associated with a cardinal direction of west?"}], "gold_sql": "SELECT col1 FROM data WHERE col0 = 'west';", "gold_answer": ["kyathabade ကြာသပတေး"], "tokens": 198, "prompt_tokens": 184, "raw_output": "SELECT col1 FROM data WHERE col0 = 'west'", "seconds": 0.16817524982616305, "generation_tokens": 13, "valid_sql": true, "answer_correct": true, "exact_sql": true, "predicted_answer": ["kyathabade ကြာသပတေး"], "answer_correct_without_fence": true}
{"source_split": "dev", "source_row": 8129, "table_id": "2-12264570-2", "question": "Which conference held at School of whiting?", "headers": ["School", "Location", "Mascot", "County", "Joined", "Previous Conference", "Left", "Conference joined"], "types": ["text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 5, "conds": [[0, 0, "whiting"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"School\"\ncol1 (text): \"Location\"\ncol2 (text): \"Mascot\"\ncol3 (text): \"County\"\ncol4 (text): \"Joined\"\ncol5 (text): \"Previous Conference\"\ncol6 (text): \"Left\"\ncol7 (text): \"Conference joined\"\n\nQuestion: Which conference held at School of whiting?"}], "gold_sql": "SELECT col5 FROM data WHERE col0 = 'whiting';", "gold_answer": ["independents"], "tokens": 207, "prompt_tokens": 192, "raw_output": "SELECT Conference joined FROM data WHERE School = 'School of Whiting'", "seconds": 0.1629724157974124, "generation_tokens": 15, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 238, "table_id": "1-1242447-1", "question": "What is the strongs words compounded when the english spelling is jonadab?", "headers": ["Strongs #", "Hebrew word", "Strongs Transliteration", "Strongs Words Compounded", "English Spelling"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[4, 0, "Jonadab"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Strongs #\"\ncol1 (text): \"Hebrew word\"\ncol2 (text): \"Strongs Transliteration\"\ncol3 (text): \"Strongs Words Compounded\"\ncol4 (text): \"English Spelling\"\n\nQuestion: What is the strongs words compounded when the english spelling is jonadab?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 = 'jonadab';", "gold_answer": ["\"a form of 3082\""], "tokens": 201, "prompt_tokens": 185, "raw_output": "SELECT col4 FROM data WHERE col3 = 'jonadab'", "seconds": 0.15843400033190846, "generation_tokens": 15, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6260, "table_id": "2-11173586-1", "question": "What sum of Attendance has a Week smaller than 10, and a Result of l 30-21?", "headers": ["Week", "Date", "Opponent", "Result", "Attendance"], "types": ["real", "text", "text", "text", "real"], "query": {"sel": 4, "conds": [[0, 2, 10], [3, 0, "l 30-21"]], "agg": 4}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Week\"\ncol1 (text): \"Date\"\ncol2 (text): \"Opponent\"\ncol3 (text): \"Result\"\ncol4 (real): \"Attendance\"\n\nQuestion: What sum of Attendance has a Week smaller than 10, and a Result of l 30-21?"}], "gold_sql": "SELECT SUM(col4) FROM data WHERE col0 < 10 AND col3 = 'l 30-21';", "gold_answer": [10.0], "tokens": 210, "prompt_tokens": 181, "raw_output": "SELECT SUM(Attendance) FROM data WHERE Week < 10 AND Result = '30-21'", "seconds": 0.20278891688212752, "generation_tokens": 24, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5020, "table_id": "2-16435448-4", "question": "What was the outcome when Jing-Jing Lu was the opponent?", "headers": ["Outcome", "Date", "Tournament", "Surface", "Opponent", "Score"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[4, 0, "jing-jing lu"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Outcome\"\ncol1 (text): \"Date\"\ncol2 (text): \"Tournament\"\ncol3 (text): \"Surface\"\ncol4 (text): \"Opponent\"\ncol5 (text): \"Score\"\n\nQuestion: What was the outcome when Jing-Jing Lu was the opponent?"}], "gold_sql": "SELECT col0 FROM data WHERE col4 = 'jing-jing lu';", "gold_answer": ["runner-up"], "tokens": 195, "prompt_tokens": 178, "raw_output": "SELECT Outcome FROM data WHERE Opponent = 'Jing-Jing Lu'", "seconds": 0.16465108329430223, "generation_tokens": 16, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1271, "table_id": "1-21734764-1", "question": "Which administrative division had a population of 2011 according to the siak database of 3,672,994?", "headers": ["Administrative division", "Area (km²) 2005", "Population 2010 Census", "Population 2011 SIAK Database", "Population density (/km² 2010)"], "types": ["text", "text", "real", "text", "real"], "query": {"sel": 0, "conds": [[3, 0, "3,672,994"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Administrative division\"\ncol1 (text): \"Area (km²) 2005\"\ncol2 (real): \"Population 2010 Census\"\ncol3 (text): \"Population 2011 SIAK Database\"\ncol4 (real): \"Population density (/km² 2010)\"\n\nQuestion: Which administrative division had a population of 2011 according to the siak database of 3,672,994?"}], "gold_sql": "SELECT col0 FROM data WHERE col3 = '3,672,994';", "gold_answer": ["bandung regency"], "tokens": 242, "prompt_tokens": 220, "raw_output": "SELECT col0 FROM data WHERE col2 = 2011 AND col4 = 3672994", "seconds": 0.21816795924678445, "generation_tokens": 27, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6565, "table_id": "2-1800293-3", "question": "What is Example Spelled Free, when Example Spelled Checked is \"op /ɔp/ (\"on\")\"?", "headers": ["Letter", "Pronunciation spelled free", "Pronunciation spelled checked", "Example spelled free", "Example spelled checked"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[4, 0, "op /ɔp/ (\"on\")"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Letter\"\ncol1 (text): \"Pronunciation spelled free\"\ncol2 (text): \"Pronunciation spelled checked\"\ncol3 (text): \"Example spelled free\"\ncol4 (text): \"Example spelled checked\"\n\nQuestion: What is Example Spelled Free, when Example Spelled Checked is \"op /ɔp/ (\"on\")\"?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 = 'op /ɔp/ (\"on\")';", "gold_answer": ["open /ˈoːpə(n)/ (\"open\")"], "tokens": 211, "prompt_tokens": 190, "raw_output": "SELECT col0 FROM data WHERE col3 = \"Example spelled free\" AND col4 = \"op /ɔp/\"", "seconds": 0.2095099575817585, "generation_tokens": 25, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 653, "table_id": "1-157826-1", "question": "What is the total population in the city/town of Arendal?", "headers": ["City/town", "Municipality", "County", "City/town status", "Population"], "types": ["text", "text", "text", "real", "real"], "query": {"sel": 4, "conds": [[0, 0, "Arendal"]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"City/town\"\ncol1 (text): \"Municipality\"\ncol2 (text): \"County\"\ncol3 (real): \"City/town status\"\ncol4 (real): \"Population\"\n\nQuestion: What is the total population in the city/town of Arendal?"}], "gold_sql": "SELECT COUNT(col4) FROM data WHERE col0 = 'arendal';", "gold_answer": [1], "tokens": 194, "prompt_tokens": 177, "raw_output": "SELECT SUM(col4) FROM data WHERE col0 = 'Arendal'", "seconds": 0.17263616668060422, "generation_tokens": 17, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [null], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 8006, "table_id": "2-1214638-4", "question": "Name the formed that has type of 4-cor", "headers": ["Units", "Type", "Period used", "Formed from", "Trailers"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[1, 0, "4-cor"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Units\"\ncol1 (text): \"Type\"\ncol2 (text): \"Period used\"\ncol3 (text): \"Formed from\"\ncol4 (text): \"Trailers\"\n\nQuestion: Name the formed that has type of 4-cor"}], "gold_sql": "SELECT col3 FROM data WHERE col1 = '4-cor';", "gold_answer": ["pan/pul/res cars"], "tokens": 184, "prompt_tokens": 169, "raw_output": "SELECT col4 FROM data WHERE col1 = \"Type\" AND col2 = \"4-cor\"", "seconds": 0.19049741700291634, "generation_tokens": 21, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2079, "table_id": "1-27539272-5", "question": "what is the maximum number of points?", "headers": ["Game", "November", "Opponent", "Score", "Location/Attendance", "Record", "Points"], "types": ["real", "real", "text", "text", "text", "text", "real"], "query": {"sel": 6, "conds": [], "agg": 1}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Game\"\ncol1 (real): \"November\"\ncol2 (text): \"Opponent\"\ncol3 (text): \"Score\"\ncol4 (text): \"Location/Attendance\"\ncol5 (text): \"Record\"\ncol6 (real): \"Points\"\n\nQuestion: what is the maximum number of points?"}], "gold_sql": "SELECT MAX(col6) FROM data;", "gold_answer": [16.0], "tokens": 192, "prompt_tokens": 182, "raw_output": "SELECT MAX(Points) FROM data", "seconds": 0.13521199999377131, "generation_tokens": 8, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7412, "table_id": "2-18782055-2", "question": "How many bronze medals were won when the total is more than 1, and gold is more than 0?", "headers": ["Rank", "Nation", "Gold", "Silver", "Bronze", "Total"], "types": ["real", "text", "real", "real", "real", "real"], "query": {"sel": 4, "conds": [[5, 1, 1], [2, 1, 0]], "agg": 5}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Rank\"\ncol1 (text): \"Nation\"\ncol2 (real): \"Gold\"\ncol3 (real): \"Silver\"\ncol4 (real): \"Bronze\"\ncol5 (real): \"Total\"\n\nQuestion: How many bronze medals were won when the total is more than 1, and gold is more than 0?"}], "gold_sql": "SELECT AVG(col4) FROM data WHERE col5 > 1 AND col2 > 0;", "gold_answer": [null], "tokens": 209, "prompt_tokens": 187, "raw_output": "SELECT COUNT(T1.col5) FROM data AS T1 INNER JOIN data AS T2 ON T1.col0 = T2.col0 WHERE T1.col4 > 1 AND T1.col3 > 0", "seconds": 0.2916355412453413, "generation_tokens": 45, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4822, "table_id": "2-1676073-12", "question": "What is Drawn, when Points Against is \"686\"?", "headers": ["Club", "Played", "Drawn", "Lost", "Points for", "Points against", "Points difference", "Points"], "types": ["text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[5, 0, "686"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Club\"\ncol1 (text): \"Played\"\ncol2 (text): \"Drawn\"\ncol3 (text): \"Lost\"\ncol4 (text): \"Points for\"\ncol5 (text): \"Points against\"\ncol6 (text): \"Points difference\"\ncol7 (text): \"Points\"\n\nQuestion: What is Drawn, when Points Against is \"686\"?"}], "gold_sql": "SELECT col2 FROM data WHERE col5 = '686';", "gold_answer": ["1"], "tokens": 213, "prompt_tokens": 197, "raw_output": "SELECT Drawn FROM data WHERE Points Against = \"686\"", "seconds": 0.17272099992260337, "generation_tokens": 15, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Invalid expression / Unexpected token. Line 1, Col: 43.\n  SELECT Drawn FROM data WHERE Points \u001b[4mAgainst\u001b[0m = \"686\"", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7128, "table_id": "2-18150723-4", "question": "What was Anders Forsbrand's score when the TO par is +4?", "headers": ["Place", "Player", "Country", "Score", "To par"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[4, 0, "+4"], [1, 0, "anders forsbrand"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Place\"\ncol1 (text): \"Player\"\ncol2 (text): \"Country\"\ncol3 (text): \"Score\"\ncol4 (text): \"To par\"\n\nQuestion: What was Anders Forsbrand's score when the TO par is +4?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 = '+4' AND col1 = 'anders forsbrand';", "gold_answer": ["71-73=144"], "tokens": 194, "prompt_tokens": 171, "raw_output": "SELECT Score FROM data WHERE Player = 'Anders Forsbrand' AND TO par = 4", "seconds": 0.1847449168562889, "generation_tokens": 20, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Invalid expression / Unexpected token. Line 1, Col: 67.\n  SELECT Score FROM data WHERE Player = 'Anders Forsbrand' AND TO \u001b[4mpar\u001b[0m = 4", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6859, "table_id": "2-17759945-8", "question": "What is the weight of the entry that has a date of birth of 1981-11-21?", "headers": ["Name", "Pos.", "Height", "Weight", "Date of Birth", "Club"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[4, 0, "1981-11-21"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"Pos.\"\ncol2 (text): \"Height\"\ncol3 (text): \"Weight\"\ncol4 (text): \"Date of Birth\"\ncol5 (text): \"Club\"\n\nQuestion: What is the weight of the entry that has a date of birth of 1981-11-21?"}], "gold_sql": "SELECT col3 FROM data WHERE col4 = '1981-11-21';", "gold_answer": ["kg (lb)"], "tokens": 214, "prompt_tokens": 191, "raw_output": "SELECT Weight FROM data WHERE Date_of_Birth = '1981-11-21'", "seconds": 0.2022770419716835, "generation_tokens": 23, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5185, "table_id": "2-17100961-26", "question": "What is the average number of rounds for billy hicks who had an overall pick number bigger than 310?", "headers": ["Round", "Pick", "Overall", "Name", "Position", "College"], "types": ["real", "real", "real", "text", "text", "text"], "query": {"sel": 0, "conds": [[2, 1, 310], [3, 0, "billy hicks"]], "agg": 5}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Round\"\ncol1 (real): \"Pick\"\ncol2 (real): \"Overall\"\ncol3 (text): \"Name\"\ncol4 (text): \"Position\"\ncol5 (text): \"College\"\n\nQuestion: What is the average number of rounds for billy hicks who had an overall pick number bigger than 310?"}], "gold_sql": "SELECT AVG(col0) FROM data WHERE col2 > 310 AND col3 = 'billy hicks';", "gold_answer": [28.0], "tokens": 215, "prompt_tokens": 188, "raw_output": "SELECT AVG(col0) FROM data WHERE col3 = 'Billy Hicks' AND col2 > 310", "seconds": 0.2067617499269545, "generation_tokens": 25, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [null], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4379, "table_id": "2-16894271-5", "question": "Which year has a 2003 of lq?", "headers": ["Tournament", "2003", "2004", "2005", "2006", "2007", "2008", "2009", "2010", "2011", "2012", "2013"], "types": ["text", "text", "text", "text", "text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 7, "conds": [[1, 0, "lq"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Tournament\"\ncol1 (text): \"2003\"\ncol2 (text): \"2004\"\ncol3 (text): \"2005\"\ncol4 (text): \"2006\"\ncol5 (text): \"2007\"\ncol6 (text): \"2008\"\ncol7 (text): \"2009\"\ncol8 (text): \"2010\"\ncol9 (text): \"2011\"\ncol10 (text): \"2012\"\ncol11 (text): \"2013\"\n\nQuestion: Which year has a 2003 of lq?"}], "gold_sql": "SELECT col7 FROM data WHERE col1 = 'lq';", "gold_answer": ["1r", "1r", "lq", "lq"], "tokens": 275, "prompt_tokens": 260, "raw_output": "SELECT col1 FROM data WHERE col0 = 'Tournament' AND col2 = '2003'", "seconds": 0.21306020906195045, "generation_tokens": 24, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2679, "table_id": "2-10747009-7", "question": "Which home team competed against the away team Geelong?", "headers": ["Home team", "Home team score", "Away team", "Away team score", "Venue", "Crowd", "Date"], "types": ["text", "text", "text", "text", "text", "real", "text"], "query": {"sel": 0, "conds": [[2, 0, "geelong"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Home team\"\ncol1 (text): \"Home team score\"\ncol2 (text): \"Away team\"\ncol3 (text): \"Away team score\"\ncol4 (text): \"Venue\"\ncol5 (real): \"Crowd\"\ncol6 (text): \"Date\"\n\nQuestion: Which home team competed against the away team Geelong?"}], "gold_sql": "SELECT col0 FROM data WHERE col2 = 'geelong';", "gold_answer": ["essendon"], "tokens": 204, "prompt_tokens": 189, "raw_output": "SELECT col0 FROM data WHERE col1 = 'Geelong' AND col2 = 'Away team'", "seconds": 0.19437629217281938, "generation_tokens": 22, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5873, "table_id": "2-11233454-2", "question": "Which award was nominated for in 2000?", "headers": ["Year", "Organisation", "Ceremony", "Award", "Result"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[0, 0, 2000], [4, 0, "nominated"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Year\"\ncol1 (text): \"Organisation\"\ncol2 (text): \"Ceremony\"\ncol3 (text): \"Award\"\ncol4 (text): \"Result\"\n\nQuestion: Which award was nominated for in 2000?"}], "gold_sql": "SELECT col3 FROM data WHERE col0 = 2000 AND col4 = 'nominated';", "gold_answer": ["best cover art", "best group", "single of the year"], "tokens": 194, "prompt_tokens": 170, "raw_output": "SELECT Award FROM data WHERE Year = 2000", "seconds": 0.15784912509843707, "generation_tokens": 13, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4933, "table_id": "2-16299161-2", "question": "Which poll source had an other of 15%?", "headers": ["Poll source", "Dates administered", "Jim DeMint (R)", "Alvin Greene (D)", "Tom Clements (G)", "Other", "Undecided"], "types": ["text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[5, 0, "15%"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Poll source\"\ncol1 (text): \"Dates administered\"\ncol2 (text): \"Jim DeMint (R)\"\ncol3 (text): \"Alvin Greene (D)\"\ncol4 (text): \"Tom Clements (G)\"\ncol5 (text): \"Other\"\ncol6 (text): \"Undecided\"\n\nQuestion: Which poll source had an other of 15%?"}], "gold_sql": "SELECT col0 FROM data WHERE col5 = '15%';", "gold_answer": ["rasmussen reports"], "tokens": 215, "prompt_tokens": 199, "raw_output": "SELECT col0 FROM data WHERE col6 = 'Undecided' AND col1 LIKE '%other%' AND col2 LIKE '%15%'", "seconds": 0.22475633304566145, "generation_tokens": 29, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5564, "table_id": "2-11554565-3", "question": "What is the record for the game on January 19?", "headers": ["Date", "Visitor", "Score", "Home", "Record"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[0, 0, "january 19"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Visitor\"\ncol2 (text): \"Score\"\ncol3 (text): \"Home\"\ncol4 (text): \"Record\"\n\nQuestion: What is the record for the game on January 19?"}], "gold_sql": "SELECT col4 FROM data WHERE col0 = 'january 19';", "gold_answer": ["8–1–0"], "tokens": 186, "prompt_tokens": 168, "raw_output": "SELECT Record FROM data WHERE col0 = 'January 19'", "seconds": 0.16623766673728824, "generation_tokens": 15, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5729, "table_id": "2-10808681-17", "question": "What did the away team score when they were playing collingwood?", "headers": ["Home team", "Home team score", "Away team", "Away team score", "Venue", "Crowd", "Date"], "types": ["text", "text", "text", "text", "text", "real", "text"], "query": {"sel": 3, "conds": [[0, 0, "collingwood"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Home team\"\ncol1 (text): \"Home team score\"\ncol2 (text): \"Away team\"\ncol3 (text): \"Away team score\"\ncol4 (text): \"Venue\"\ncol5 (real): \"Crowd\"\ncol6 (text): \"Date\"\n\nQuestion: What did the away team score when they were playing collingwood?"}], "gold_sql": "SELECT col3 FROM data WHERE col0 = 'collingwood';", "gold_answer": ["8.12 (60)"], "tokens": 208, "prompt_tokens": 192, "raw_output": "SELECT col2 FROM data WHERE col0 = 'Collingwood'", "seconds": 0.16682066721841693, "generation_tokens": 15, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7349, "table_id": "2-18964684-2", "question": "What years did the sdrive16d model have a Torque of n·m (lb·ft)@1750-3000?", "headers": ["Model", "Years", "Engine code", "Power", "Torque"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[4, 0, "n·m (lb·ft)@1750-3000"], [0, 0, "sdrive16d"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Model\"\ncol1 (text): \"Years\"\ncol2 (text): \"Engine code\"\ncol3 (text): \"Power\"\ncol4 (text): \"Torque\"\n\nQuestion: What years did the sdrive16d model have a Torque of n·m (lb·ft)@1750-3000?"}], "gold_sql": "SELECT col1 FROM data WHERE col4 = 'n·m (lb·ft)@1750-3000' AND col0 = 'sdrive16d';", "gold_answer": ["2012-"], "tokens": 233, "prompt_tokens": 191, "raw_output": "SELECT col1 FROM data WHERE col0 = 'sdrive16d' AND col4 = 'n·m (lb·ft)@1750-3000'", "seconds": 0.27049724990502, "generation_tokens": 41, "valid_sql": true, "answer_correct": true, "exact_sql": false, "predicted_answer": ["2012-"], "answer_correct_without_fence": true}
{"source_split": "dev", "source_row": 4347, "table_id": "2-1662630-5", "question": "what is the sum of wins when events is 13 and top-5 is less than 1?", "headers": ["Tournament", "Wins", "Top-5", "Top-10", "Top-25", "Events", "Cuts made"], "types": ["text", "real", "real", "real", "real", "real", "real"], "query": {"sel": 1, "conds": [[5, 0, 13], [2, 2, 1]], "agg": 4}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Tournament\"\ncol1 (real): \"Wins\"\ncol2 (real): \"Top-5\"\ncol3 (real): \"Top-10\"\ncol4 (real): \"Top-25\"\ncol5 (real): \"Events\"\ncol6 (real): \"Cuts made\"\n\nQuestion: what is the sum of wins when events is 13 and top-5 is less than 1?"}], "gold_sql": "SELECT SUM(col1) FROM data WHERE col5 = 13 AND col2 < 1;", "gold_answer": [null], "tokens": 227, "prompt_tokens": 204, "raw_output": "SELECT SUM(wins) FROM data WHERE events = 13 AND top-5 < 1", "seconds": 0.19026133324950933, "generation_tokens": 21, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6149, "table_id": "2-10824095-2", "question": "Which home team played the away team of collingwood?", "headers": ["Home team", "Home team score", "Away team", "Away team score", "Venue", "Crowd", "Date"], "types": ["text", "text", "text", "text", "text", "real", "text"], "query": {"sel": 0, "conds": [[2, 0, "collingwood"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Home team\"\ncol1 (text): \"Home team score\"\ncol2 (text): \"Away team\"\ncol3 (text): \"Away team score\"\ncol4 (text): \"Venue\"\ncol5 (real): \"Crowd\"\ncol6 (text): \"Date\"\n\nQuestion: Which home team played the away team of collingwood?"}], "gold_sql": "SELECT col0 FROM data WHERE col2 = 'collingwood';", "gold_answer": ["footscray"], "tokens": 206, "prompt_tokens": 190, "raw_output": "SELECT col0 FROM data WHERE col1 = 'Collingwood' AND col2 = 'Away team'", "seconds": 0.1873694579117, "generation_tokens": 23, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6986, "table_id": "2-17679324-9", "question": "What is the surface of the tournament with a runner-up outcome and dudi sela as the opponent?", "headers": ["Outcome", "Date", "Tournament", "Surface", "Opponent", "Score"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[0, 0, "runner-up"], [4, 0, "dudi sela"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Outcome\"\ncol1 (text): \"Date\"\ncol2 (text): \"Tournament\"\ncol3 (text): \"Surface\"\ncol4 (text): \"Opponent\"\ncol5 (text): \"Score\"\n\nQuestion: What is the surface of the tournament with a runner-up outcome and dudi sela as the opponent?"}], "gold_sql": "SELECT col3 FROM data WHERE col0 = 'runner-up' AND col4 = 'dudi sela';", "gold_answer": ["grass"], "tokens": 211, "prompt_tokens": 186, "raw_output": "SELECT Surface FROM data WHERE Outcome = 'Runner-up' AND Opponent = 'Dudi Sela'", "seconds": 0.18267329223453999, "generation_tokens": 22, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7623, "table_id": "2-18874135-1", "question": "Who is the Winner of the Nissan Motorsport Australia Team at the Oran Park Raceway?", "headers": ["Date", "Series", "Circuit", "City / State", "Winner", "Team"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[5, 0, "nissan motorsport australia"], [2, 0, "oran park raceway"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Series\"\ncol2 (text): \"Circuit\"\ncol3 (text): \"City / State\"\ncol4 (text): \"Winner\"\ncol5 (text): \"Team\"\n\nQuestion: Who is the Winner of the Nissan Motorsport Australia Team at the Oran Park Raceway?"}], "gold_sql": "SELECT col4 FROM data WHERE col5 = 'nissan motorsport australia' AND col2 = 'oran park raceway';", "gold_answer": ["mark skaife"], "tokens": 213, "prompt_tokens": 185, "raw_output": "SELECT Winner FROM data WHERE Series = 'Nissan Motorsport Australia Team' AND Circuit = 'Oran Park Raceway'", "seconds": 0.2002845830284059, "generation_tokens": 26, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2317, "table_id": "1-29218221-2", "question": "Who was the rider with a Fri 3 June time of 18' 19.68 123.516mph?", "headers": ["Rank", "Rider", "Mon 30 May", "Tues 31 May", "Wed 1 June", "Thurs 2 June", "Fri 3 June"], "types": ["real", "text", "text", "text", "text", "text", "text"], "query": {"sel": 1, "conds": [[6, 0, "18' 19.68 123.516mph"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Rank\"\ncol1 (text): \"Rider\"\ncol2 (text): \"Mon 30 May\"\ncol3 (text): \"Tues 31 May\"\ncol4 (text): \"Wed 1 June\"\ncol5 (text): \"Thurs 2 June\"\ncol6 (text): \"Fri 3 June\"\n\nQuestion: Who was the rider with a Fri 3 June time of 18' 19.68 123.516mph?"}], "gold_sql": "SELECT col1 FROM data WHERE col6 = '18'' 19.68 123.516mph';", "gold_answer": ["john mcguinness 599cc honda"], "tokens": 254, "prompt_tokens": 223, "raw_output": "SELECT rider FROM data WHERE mon 30 May = '18' AND tues 31 May = '19.68' AND wed 1 June = '123.516mph'", "seconds": 0.3006086670793593, "generation_tokens": 47, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Invalid expression / Unexpected token. Line 1, Col: 35.\n  SELECT rider FROM data WHERE mon \u001b[4m30\u001b[0m May = '18' AND tues 31 May = '19.68' AND wed 1 June = '123.516mph'", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1418, "table_id": "1-22993636-5", "question": "What is the lowest number of 3 pointers that occured in games with 52 steals?", "headers": ["Player", "Games Played", "Minutes", "Field Goals", "Three Pointers", "Free Throws", "Rebounds", "Assists", "Blocks", "Steals", "Points"], "types": ["text", "real", "real", "real", "real", "real", "real", "real", "real", "real", "real"], "query": {"sel": 4, "conds": [[9, 0, 52]], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Player\"\ncol1 (real): \"Games Played\"\ncol2 (real): \"Minutes\"\ncol3 (real): \"Field Goals\"\ncol4 (real): \"Three Pointers\"\ncol5 (real): \"Free Throws\"\ncol6 (real): \"Rebounds\"\ncol7 (real): \"Assists\"\ncol8 (real): \"Blocks\"\ncol9 (real): \"Steals\"\ncol10 (real): \"Points\"\n\nQuestion: What is the lowest number of 3 pointers that occured in games with 52 steals?"}], "gold_sql": "SELECT MIN(col4) FROM data WHERE col9 = 52;", "gold_answer": [80.0], "tokens": 248, "prompt_tokens": 231, "raw_output": "SELECT MIN(col3) FROM data WHERE col8 = 52", "seconds": 0.1629794160835445, "generation_tokens": 15, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [null], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5788, "table_id": "2-1121500-1", "question": "What is the result in oakland?", "headers": ["Goal", "Date", "Venue", "Result", "Competition"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[2, 0, "oakland"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Goal\"\ncol1 (text): \"Date\"\ncol2 (text): \"Venue\"\ncol3 (text): \"Result\"\ncol4 (text): \"Competition\"\n\nQuestion: What is the result in oakland?"}], "gold_sql": "SELECT col3 FROM data WHERE col2 = 'oakland';", "gold_answer": ["7–2", "7–2", "7–2", "7–2"], "tokens": 180, "prompt_tokens": 164, "raw_output": "SELECT Result FROM data WHERE Venue = 'Oakland'", "seconds": 0.15002741711214185, "generation_tokens": 12, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6611, "table_id": "2-1824778-2", "question": "What payload was on November 26, 2011, 15:02?", "headers": ["Date and time( UTC )", "Type", "Serial no.", "Launch site", "Payload", "Type of payload", "Orbit", "Outcome"], "types": ["text", "real", "text", "text", "text", "text", "text", "text"], "query": {"sel": 5, "conds": [[0, 0, "november 26, 2011, 15:02"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date and time( UTC )\"\ncol1 (real): \"Type\"\ncol2 (text): \"Serial no.\"\ncol3 (text): \"Launch site\"\ncol4 (text): \"Payload\"\ncol5 (text): \"Type of payload\"\ncol6 (text): \"Orbit\"\ncol7 (text): \"Outcome\"\n\nQuestion: What payload was on November 26, 2011, 15:02?"}], "gold_sql": "SELECT col5 FROM data WHERE col0 = 'november 26, 2011, 15:02';", "gold_answer": ["mars rover"], "tokens": 242, "prompt_tokens": 211, "raw_output": "SELECT payload FROM data WHERE launch_site = 'November 26, 2011' AND type = 'Payload' AND outcome = 'Success' AND col0 = '2011-11-26 15:02'", "seconds": 0.34432970918715, "generation_tokens": 55, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5118, "table_id": "2-17120059-4", "question": "What is the Result of the Game on April 9?", "headers": ["Game", "Date", "Home Team", "Result", "Road Team"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[1, 0, "april 9"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Game\"\ncol1 (text): \"Date\"\ncol2 (text): \"Home Team\"\ncol3 (text): \"Result\"\ncol4 (text): \"Road Team\"\n\nQuestion: What is the Result of the Game on April 9?"}], "gold_sql": "SELECT col3 FROM data WHERE col1 = 'april 9';", "gold_answer": ["100-102"], "tokens": 186, "prompt_tokens": 169, "raw_output": "SELECT Result FROM data WHERE Date = 'April 9'", "seconds": 0.16586587484925985, "generation_tokens": 13, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 2347, "table_id": "1-29612224-1", "question": "How many different combinations of team colors are there in all the schools in Maroa, Illinois?", "headers": ["Institution", "Location", "Mascot", "Colors", "Affiliation", "9–12 enrollment", "School website"], "types": ["text", "text", "text", "text", "text", "real", "text"], "query": {"sel": 3, "conds": [[1, 0, "Maroa, Illinois"]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Institution\"\ncol1 (text): \"Location\"\ncol2 (text): \"Mascot\"\ncol3 (text): \"Colors\"\ncol4 (text): \"Affiliation\"\ncol5 (real): \"9–12 enrollment\"\ncol6 (text): \"School website\"\n\nQuestion: How many different combinations of team colors are there in all the schools in Maroa, Illinois?"}], "gold_sql": "SELECT COUNT(col3) FROM data WHERE col1 = 'maroa, illinois';", "gold_answer": [1], "tokens": 219, "prompt_tokens": 199, "raw_output": "SELECT COUNT(DISTINCT Colors) FROM data WHERE Location = 'Maroa, Illinois'", "seconds": 0.18827795889228582, "generation_tokens": 19, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5138, "table_id": "2-166170-5", "question": "who is the opponent on 18 january 1993?", "headers": ["Outcome", "Date", "Championship", "Surface", "Opponent", "Score"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[1, 0, "18 january 1993"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Outcome\"\ncol1 (text): \"Date\"\ncol2 (text): \"Championship\"\ncol3 (text): \"Surface\"\ncol4 (text): \"Opponent\"\ncol5 (text): \"Score\"\n\nQuestion: who is the opponent on 18 january 1993?"}], "gold_sql": "SELECT col4 FROM data WHERE col1 = '18 january 1993';", "gold_answer": ["pete sampras"], "tokens": 203, "prompt_tokens": 181, "raw_output": "SELECT Opponent FROM data WHERE Date = '18 january 1993' AND Championship = '1993'", "seconds": 0.2215971671976149, "generation_tokens": 29, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 8220, "table_id": "2-12206431-2", "question": "What team did they lose to when they had a 28-22 record?", "headers": ["Date", "Opponent", "Score", "Loss", "Attendance", "Record"], "types": ["text", "text", "text", "text", "text", "text"], "query": {"sel": 3, "conds": [[5, 0, "28-22"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Date\"\ncol1 (text): \"Opponent\"\ncol2 (text): \"Score\"\ncol3 (text): \"Loss\"\ncol4 (text): \"Attendance\"\ncol5 (text): \"Record\"\n\nQuestion: What team did they lose to when they had a 28-22 record?"}], "gold_sql": "SELECT col3 FROM data WHERE col5 = '28-22';", "gold_answer": ["mohler (0-1)"], "tokens": 200, "prompt_tokens": 182, "raw_output": "SELECT Opponent FROM data WHERE Record = 28 AND Loss = 22", "seconds": 0.1831402499228716, "generation_tokens": 18, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4658, "table_id": "2-16845107-1", "question": "Which Title has a Type of album and a Year larger than 1986?", "headers": ["Year", "Artist", "Title", "Label", "Type"], "types": ["real", "text", "text", "text", "text"], "query": {"sel": 2, "conds": [[4, 0, "album"], [0, 1, 1986]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Year\"\ncol1 (text): \"Artist\"\ncol2 (text): \"Title\"\ncol3 (text): \"Label\"\ncol4 (text): \"Type\"\n\nQuestion: Which Title has a Type of album and a Year larger than 1986?"}], "gold_sql": "SELECT col2 FROM data WHERE col4 = 'album' AND col0 > 1986;", "gold_answer": ["g tom mac", "the immortal (original soundtrack)", "sunset junction (original soundtrack)", "thou shalt not fall"], "tokens": 197, "prompt_tokens": 173, "raw_output": "SELECT Title FROM data WHERE Type = 'album' AND Year > 1986", "seconds": 0.178551624994725, "generation_tokens": 19, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7448, "table_id": "2-18771190-15", "question": "What's the lost when there were more than 16 points and had a drawn less than 1?", "headers": ["Position", "Name", "Played", "Drawn", "Lost", "Points"], "types": ["real", "text", "real", "real", "real", "real"], "query": {"sel": 4, "conds": [[5, 1, 16], [3, 2, 1]], "agg": 4}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Position\"\ncol1 (text): \"Name\"\ncol2 (real): \"Played\"\ncol3 (real): \"Drawn\"\ncol4 (real): \"Lost\"\ncol5 (real): \"Points\"\n\nQuestion: What's the lost when there were more than 16 points and had a drawn less than 1?"}], "gold_sql": "SELECT SUM(col4) FROM data WHERE col5 > 16 AND col3 < 1;", "gold_answer": [null], "tokens": 209, "prompt_tokens": 186, "raw_output": "SELECT Lost FROM data WHERE Played > 16 AND Drawn < 1", "seconds": 0.17235816596075892, "generation_tokens": 17, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4068, "table_id": "2-15625088-7", "question": "Which Tie no has a Score of 0–1, and a Date of 9 march 1953?", "headers": ["Tie no", "Home team", "Score", "Away team", "Date"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 0, "conds": [[2, 0, "0–1"], [4, 0, "9 march 1953"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Tie no\"\ncol1 (text): \"Home team\"\ncol2 (text): \"Score\"\ncol3 (text): \"Away team\"\ncol4 (text): \"Date\"\n\nQuestion: Which Tie no has a Score of 0–1, and a Date of 9 march 1953?"}], "gold_sql": "SELECT col0 FROM data WHERE col2 = '0–1' AND col4 = '9 march 1953';", "gold_answer": ["replay"], "tokens": 213, "prompt_tokens": 184, "raw_output": "SELECT col0 FROM data WHERE col2 = '0-1' AND col3 = '9 march 1953'", "seconds": 0.2154449583031237, "generation_tokens": 28, "valid_sql": true, "answer_correct": false, "exact_sql": false, "predicted_answer": [], "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6414, "table_id": "2-12005739-2", "question": "Can you tell me the average Points that has a Pts/game larger than 4, and the Nation of england, and the Games smaller than 5?", "headers": ["Name", "Nation", "Points", "Games", "Pts/game"], "types": ["text", "text", "real", "real", "real"], "query": {"sel": 2, "conds": [[4, 1, 4], [1, 0, "england"], [3, 2, 5]], "agg": 5}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"Nation\"\ncol2 (real): \"Points\"\ncol3 (real): \"Games\"\ncol4 (real): \"Pts/game\"\n\nQuestion: Can you tell me the average Points that has a Pts/game larger than 4, and the Nation of england, and the Games smaller than 5?"}], "gold_sql": "SELECT AVG(col2) FROM data WHERE col4 > 4 AND col1 = 'england' AND col3 < 5;", "gold_answer": [10.0], "tokens": 219, "prompt_tokens": 189, "raw_output": "SELECT AVG(T2.Points) FROM data AS T1 INNER JOIN nation AS T2 ON T1.Nation = T2.Nation WHERE T1.Pts/game > 4 AND T2.Nation = 'England' AND T1.Games < 5", "seconds": 0.3357509998604655, "generation_tokens": 54, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 3832, "table_id": "2-1576352-12", "question": "When did Demetrios III begin to hold power?", "headers": ["Type", "Name", "Title", "Royal house", "From"], "types": ["text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[1, 0, "demetrios iii"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Type\"\ncol1 (text): \"Name\"\ncol2 (text): \"Title\"\ncol3 (text): \"Royal house\"\ncol4 (text): \"From\"\n\nQuestion: When did Demetrios III begin to hold power?"}], "gold_sql": "SELECT col4 FROM data WHERE col1 = 'demetrios iii';", "gold_answer": ["100 bc"], "tokens": 184, "prompt_tokens": 167, "raw_output": "SELECT \"From\" FROM data WHERE \"Type\" = \"Demetrios III\" AND \"Royal house\" = \"Yes\"", "seconds": 0.2213922911323607, "generation_tokens": 27, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 4563, "table_id": "2-17409365-3", "question": "What is the points when the try bonus is 1?", "headers": ["Club", "Played", "Drawn", "Lost", "Points for", "Points against", "Tries for", "Tries against", "Try bonus"], "types": ["text", "text", "text", "text", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[8, 0, "1"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Club\"\ncol1 (text): \"Played\"\ncol2 (text): \"Drawn\"\ncol3 (text): \"Lost\"\ncol4 (text): \"Points for\"\ncol5 (text): \"Points against\"\ncol6 (text): \"Tries for\"\ncol7 (text): \"Tries against\"\ncol8 (text): \"Try bonus\"\n\nQuestion: What is the points when the try bonus is 1?"}], "gold_sql": "SELECT col4 FROM data WHERE col8 = '1';", "gold_answer": ["219"], "tokens": 221, "prompt_tokens": 207, "raw_output": "SELECT Points_for FROM data WHERE Try_bonus = 1", "seconds": 0.16305720806121826, "generation_tokens": 12, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5653, "table_id": "2-11404731-2", "question": "What's the record after week 12 with a game site of bye?", "headers": ["Week", "Date", "Opponent", "Result", "Record", "Game site"], "types": ["real", "text", "text", "text", "text", "text"], "query": {"sel": 4, "conds": [[0, 1, 12], [5, 0, "bye"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Week\"\ncol1 (text): \"Date\"\ncol2 (text): \"Opponent\"\ncol3 (text): \"Result\"\ncol4 (text): \"Record\"\ncol5 (text): \"Game site\"\n\nQuestion: What's the record after week 12 with a game site of bye?"}], "gold_sql": "SELECT col4 FROM data WHERE col0 > 12 AND col5 = 'bye';", "gold_answer": ["bye"], "tokens": 202, "prompt_tokens": 181, "raw_output": "SELECT Record FROM data WHERE Week = 12 AND Game site = 'bye'", "seconds": 0.1787868752144277, "generation_tokens": 18, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Invalid expression / Unexpected token. Line 1, Col: 53.\n  SELECT Record FROM data WHERE Week = 12 AND Game \u001b[4msite\u001b[0m = 'bye'", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 1551, "table_id": "1-2401326-1", "question": "What's percentage voted for Busg in the county where Kerry got 37.6%?", "headers": ["County", "Kerry%", "Kerry#", "Bush%", "Bush#", "Others%", "Others#"], "types": ["text", "text", "real", "text", "real", "text", "real"], "query": {"sel": 3, "conds": [[1, 0, "37.6%"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"County\"\ncol1 (text): \"Kerry%\"\ncol2 (real): \"Kerry#\"\ncol3 (text): \"Bush%\"\ncol4 (real): \"Bush#\"\ncol5 (text): \"Others%\"\ncol6 (real): \"Others#\"\n\nQuestion: What's percentage voted for Busg in the county where Kerry got 37.6%?"}], "gold_sql": "SELECT col3 FROM data WHERE col1 = '37.6%';", "gold_answer": ["60.5%"], "tokens": 214, "prompt_tokens": 196, "raw_output": "SELECT CAST(SUM(CASE WHEN col2 = 'Bush%' THEN 1 ELSE 0 END) AS REAL) * 100 / COUNT(*) FROM data WHERE col0 = 'Kerry%' AND col1 = '37.6%'", "seconds": 0.33678554091602564, "generation_tokens": 53, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 6585, "table_id": "2-1786641-1", "question": "What is the seat percentage when vote percentage is 2.4% (-8.3)?", "headers": ["Party", "Party-list votes", "Vote percentage", "Total seats", "Seat percentage"], "types": ["text", "real", "text", "text", "text"], "query": {"sel": 4, "conds": [[2, 0, "2.4% (-8.3)"]], "agg": 0}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Party\"\ncol1 (real): \"Party-list votes\"\ncol2 (text): \"Vote percentage\"\ncol3 (text): \"Total seats\"\ncol4 (text): \"Seat percentage\"\n\nQuestion: What is the seat percentage when vote percentage is 2.4% (-8.3)?"}], "gold_sql": "SELECT col4 FROM data WHERE col2 = '2.4% (-8.3)';", "gold_answer": ["0.0%"], "tokens": 201, "prompt_tokens": 179, "raw_output": "SELECT T2.seat_percentage FROM data AS T1 INNER JOIN ( SELECT col0, MAX(col1) AS max_votes FROM data GROUP BY col0 ) AS T2 ON T1.col2 = T2.col0 WHERE T1.vote_percentage = 2.4 AND T2.max_votes = (-8.3)", "seconds": 0.39896320877596736, "generation_tokens": 67, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7262, "table_id": "2-18569105-3", "question": "How many total Time listings have a 0.209 React entry and a Rank that is greater than 7?", "headers": ["Rank", "Lane", "Athlete", "Nationality", "Time", "React"], "types": ["real", "real", "text", "text", "real", "real"], "query": {"sel": 4, "conds": [[5, 0, 0.209], [0, 1, 7]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Rank\"\ncol1 (real): \"Lane\"\ncol2 (text): \"Athlete\"\ncol3 (text): \"Nationality\"\ncol4 (real): \"Time\"\ncol5 (real): \"React\"\n\nQuestion: How many total Time listings have a 0.209 React entry and a Rank that is greater than 7?"}], "gold_sql": "SELECT COUNT(col4) FROM data WHERE col5 = 0.209 AND col0 > 7;", "gold_answer": [0], "tokens": 216, "prompt_tokens": 190, "raw_output": "SELECT COUNT(*) FROM data WHERE Time = 0.209 AND Rank > 7", "seconds": 0.20117995887994766, "generation_tokens": 20, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 470, "table_id": "1-14342592-8", "question": "Name the most touchdowns for norcross", "headers": ["Player", "Position", "Starter", "Touchdowns", "Extra points", "Field goals", "Points"], "types": ["text", "text", "text", "real", "real", "real", "real"], "query": {"sel": 3, "conds": [[0, 0, "Norcross"]], "agg": 1}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Player\"\ncol1 (text): \"Position\"\ncol2 (text): \"Starter\"\ncol3 (real): \"Touchdowns\"\ncol4 (real): \"Extra points\"\ncol5 (real): \"Field goals\"\ncol6 (real): \"Points\"\n\nQuestion: Name the most touchdowns for norcross"}], "gold_sql": "SELECT MAX(col3) FROM data WHERE col0 = 'norcross';", "gold_answer": [5.0], "tokens": 200, "prompt_tokens": 183, "raw_output": "SELECT col0 FROM data WHERE col1 = \"Norcross\" ORDER BY col3 DESC LIMIT 1", "seconds": 0.19834574963897467, "generation_tokens": 22, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 443, "table_id": "1-14181578-1", "question": "what is the minimum games played with goals for/against being 7-5", "headers": ["Position", "Club (City/Town)", "Games Played", "W-L-D", "Goals For/Against", "Points"], "types": ["real", "text", "real", "text", "text", "real"], "query": {"sel": 2, "conds": [[4, 0, "7-5"]], "agg": 2}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (real): \"Position\"\ncol1 (text): \"Club (City/Town)\"\ncol2 (real): \"Games Played\"\ncol3 (text): \"W-L-D\"\ncol4 (text): \"Goals For/Against\"\ncol5 (real): \"Points\"\n\nQuestion: what is the minimum games played with goals for/against being 7-5"}], "gold_sql": "SELECT MIN(col2) FROM data WHERE col4 = '7-5';", "gold_answer": [5.0], "tokens": 211, "prompt_tokens": 193, "raw_output": "SELECT MIN(Games Played) FROM data WHERE Goals For/Against = 7 AND Goals Against = 5", "seconds": 0.19512362498790026, "generation_tokens": 23, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Expecting ). Line 1, Col: 23.\n  SELECT MIN(Games \u001b[4mPlayed\u001b[0m) FROM data WHERE Goals For/Against = 7 AND Goals Against = 5", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 5391, "table_id": "2-17309500-1", "question": "How many blocks are there when the rebounds are fewer than 5.2?", "headers": ["Season", "Points", "Rebounds", "Assists", "Blocks", "Steals", "Turnovers"], "types": ["text", "real", "real", "real", "real", "real", "real"], "query": {"sel": 4, "conds": [[2, 2, 5.2]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Season\"\ncol1 (real): \"Points\"\ncol2 (real): \"Rebounds\"\ncol3 (real): \"Assists\"\ncol4 (real): \"Blocks\"\ncol5 (real): \"Steals\"\ncol6 (real): \"Turnovers\"\n\nQuestion: How many blocks are there when the rebounds are fewer than 5.2?"}], "gold_sql": "SELECT COUNT(col4) FROM data WHERE col2 < 5.2;", "gold_answer": [0], "tokens": 209, "prompt_tokens": 191, "raw_output": "SELECT SUM(blocks) FROM data WHERE rebounds < 5.2", "seconds": 0.15951545909047127, "generation_tokens": 14, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Unknown or qualified column", "answer_correct_without_fence": false}
{"source_split": "dev", "source_row": 7961, "table_id": "2-13341768-56", "question": "Avg/G that has a GP-GS of 13–13, and a Effic smaller than 114.23 has what total of numbers?", "headers": ["Name", "GP-GS", "Effic", "Att-Cmp-Int", "Avg/G"], "types": ["text", "text", "real", "text", "real"], "query": {"sel": 4, "conds": [[1, 0, "13–13"], [2, 2, 114.23]], "agg": 3}, "messages": [{"role": "system", "content": "Translate the question into one SQLite SELECT query. Use only the table data and the column identifiers col0, col1, etc. shown below. Select one column, optionally using MAX, MIN, COUNT, SUM, or AVG. Conditions may use =, >, or < joined by AND. Do not use joins, sorting, grouping, or LIMIT. Text values in the database are lowercase; use lowercase string literals. Return only SQL, with no explanation or Markdown."}, {"role": "user", "content": "Table: data\nColumns:\ncol0 (text): \"Name\"\ncol1 (text): \"GP-GS\"\ncol2 (real): \"Effic\"\ncol3 (text): \"Att-Cmp-Int\"\ncol4 (real): \"Avg/G\"\n\nQuestion: Avg/G that has a GP-GS of 13–13, and a Effic smaller than 114.23 has what total of numbers?"}], "gold_sql": "SELECT COUNT(col4) FROM data WHERE col1 = '13–13' AND col2 < 114.23;", "gold_answer": [1], "tokens": 230, "prompt_tokens": 198, "raw_output": "SELECT AVG(GP-GS) FROM data WHERE GP-GS = 13 AND Effic < 114.23", "seconds": 0.2269964162260294, "generation_tokens": 29, "valid_sql": false, "answer_correct": false, "exact_sql": false, "error": "Query is outside WikiSQL's single-table grammar", "answer_correct_without_fence": false}