Sixteen columns are tagged "categorical" in the catalog metadata I pulled, and the counts below are the catalog's approximate distinct counts (`approx_distinct_count`). The exact `distinct_count` field was blank for every column. **True category-style columns** - `Surface` has about 4 distinct values in each of the three fact tables (`raw_kaggle`, `return_kaggle`, `serve_kaggle`). - `Rd` (round) has about 14 in each of those three tables. - `return_kaggle.against` has 1, so it carries no information. - `Name` has about 491 in each fact table. `players_tournament_man_.year` has 24, covering 2002 to 2024. - `serve_kaggle.Bpsvd` has about 236, and `serve_kaggle.Df%` has about 231. Both are stored as text. **Numeric measures that the profiler tagged "categorical" because they have few distinct values** - `raw_kaggle.Aces` has about 54. - `raw_kaggle.DFs` has about 29. - `raw_kaggle.vA` has about 62. - `raw_kaggle.1SP` has about 189. - `raw_kaggle.2SP` has about 134. **Caveats** - `Name` shows about 491 distinct values, but `players_man_` has only 462 rows. The approximate count may be slightly off, or some names may appear only in other tables. - Many other columns, such as `Tournament` (about 3,963 distinct) and `Score` (about 10,386), are tagged "classifier" or "text" rather than categorical, so I left them out. They are high-cardinality labels.