**Size:** The dataset has 5 tables with **718,470 rows** in total (step-0 row counts): - `raw_kaggle`: 237,205 rows - `return_kaggle`: 237,196 rows - `serve_kaggle`: 237,185 rows - `players_tournament_man_`: 6,422 rows - `players_man_`: 462 rows **Columns:** The two sources I have disagree, so I can't give one firm total. - The dataset card lists 21, 21, 21, 8 and 7 columns for the five tables, which adds up to **78**. - The catalog overview (step-0) reports 17, 17, 17, 4 and 3, which adds up to **38**. - The card figure is probably the complete count. The overview may be counting only the columns it profiled. I did not run `inspect_columns` on each table to settle it. **Time period:** Yes, the data is time-stamped, and it covers about **23 years, from August 2001 to October 2024**. - `Date` is stored as text in all three match-level tables. After parsing, every row converted successfully (step-3). All three tables, `raw_kaggle`, `return_kaggle` and `serve_kaggle`, span **2001-08-13 to 2024-10-28**. - The card shows narrower ranges, for example 2013–2024 for raw and return and 2006–2024 for serve. Those came from sorting the text values alphabetically, so the parsed dates are the reliable ones. - `players_tournament_man_.year` runs from **2002 to 2024** (step-2). - `players_man_.playing_years` is a text list of years per player, so it has no simple min or max. In short, this is about 718k rows, mostly one row per player per match, covering men's tennis from 2001 to 2024.