Ádám Lippai
Ádám Lippai
@Tmonster @Tishj Is this something I could take a stab at? Any directions (if to make `scan_parquet()` smarter or translate it to `read_parquet()` in simple cases) you'd prefer?
This could work indeed
Is this connected to the new numpy string dtype?
@alexander-beedie this happens with group by as well. Maybe I made the example too minimal 😅
Let me explore this more: 1. The query `SELECT count(key) AS n FROM df HAVING n >1` supposed to work, it works in the SQL engines I tested 2. I...
Agree that raising is a good temp solution, that nuance can be discussed in a separate issue (although I'd be surprised if this would be a question of flavor and...
While experimenting with the above I came across a slightly related edge-case. Let me know if I should open a separate issue for it: ```python import polars as pl pl.sql("select...
@alexander-beedie https://github.com/pola-rs/polars/issues/19310#issuecomment-2424188378 this sounds to be a common bug (regardless of the HAVING without a GROUP BY which is handled now). The aliases in aggregate context won’t work: 1. When...
@alexander-beedie do you think any of the 3 bugs could be a good first issue to solve? If you point me into the right direction I could give it a...
@alexander-beedie I might have some free time during the holidays to play with this. Can you give me any pointers where to start this from?