Implement Parquet Project Pushdown

[nils-braun]

Dask already implements the logic, to only extract columns from a parquet files which are actually needed - reducing the amount of IO.
We should make sure this feature is correctly used in dask-sql. We first need a test for it and will then need to iterate on it.

[subkanthi]

Hi @nils-braun/ @charlesbluca , is this something I can work on?

[charlesbluca]

Hi @subkanthi - I don't see any PRs in progress working on this so any help here would be greatly appreciated 🙂

[subkanthi]

Hi @subkanthi - I don't see any PRs in progress working on this so any help here would be greatly appreciated 🙂

Thanks @charlesbluca , trying to understand the process here , also making sure no one else is working on this in parallel.

[charlesbluca]

Sure - hopefully in the future, we will have a more robust standard for keeping track of what is being worked on or not. For now, I can assign this issue to you as you work on this. Feel free to let me know if you're unable to work on it anymore!