```json { "action": "final", "content": "## Table Tennis Dataset Columns\n\nThis dataset contains **22 columns** tracking table tennis match results from June-July 2022. Here's what each column represents:\n\n### Match Identification\n- **X** (BIGINT) — Row identifier/index, ranging from 1 to 7,851\n\n### Match Timing\n- **Date** (VARCHAR) — Match date, ranging from 6/10/2022 to 7/8/2022 (29 distinct dates)\n- **Time** (VARCHAR) — Match time, ranging from 0:00:00 to 9:55:00 (251 distinct times)\n\n### Players\n- **Player1** (VARCHAR) — First player's name (~635 distinct players)\n- **Player2** (VARCHAR) — Second player's name (~635 distinct players)\n\n### Match Outcome\n- **Sets_P1** (BIGINT) — Number of sets won by Player 1 (0-3)\n- **Sets_P2** (BIGINT) — Number of sets won by Player 2 (0-3)\n- **HomeWinner** (BIGINT) — Binary indicator (0 or 1) showing which player won\n\n### Game-by-Game Scores\nEach match can have up to 5 games. The columns track points scored by each player in each game:\n\n- **P1_G1** (BIGINT) — Player 1's score in Game 1 (0-21 points)\n- **P2_G1** (BIGINT) — Player 2's score in Game 1 (0-23 points)\n- **P1_G2** (BIGINT) — Player 1's score in Game 2 (0-26 points)\n- **P2_G2** (BIGINT) — Player 2's score in Game 2 (0-24 points)\n- **P1_G3** (BIGINT) — Player 1's score in Game 3 (0-22 points)\n- **P2_G3** (BIGINT) — Player 2's score in Game 3 (0-20 points)\n- **P1_G4** (BIGINT) — Player 1's score in Game 4 (0-24 points, 2,403 nulls when game not played)\n- **P2_G4** (BIGINT) — Player 2's score in Game 4 (0-22 points, 2,403 nulls when game not played)\n- **P1_G5** (BIGINT) — Player 1's score in Game 5 (0-20 points, 5,333 nulls when game not played)\n- **P2_G5** (BIGINT) — Player 2's score in Game 5 (1-22 points, 5,333 nulls when game not played)\n\n**Note:** Games 4 and 5 have many null values because most matches end in 3 games or fewer (best-of-5 format).\n\n### System Metadata\n- **_ingestion_timestamp** (TIMESTAMP) — When the data was loaded (2026-08-11 00:26:52)\n- **_batch_id** (VARCHAR) — Data ingestion batch identifier\n- **_source_file** (VARCHAR) — Original CSV file path\n- **_source_system** (VARCHAR) — Source system (\"api\")\n\nThe dataset contains **7,851 matches** with complete coverage (no nulls) for player names, dates, times, and the first 3 games of each match." } ```