Corey Bilot

Let me see if I can help. With code translations that are large, I strongly suggest using an approach that uses a left join rather than generating long `when/then/when/then ...`...

I should note: the above code *replaces* the column with new values, but the same approach using a left join should work for mapping values for new columns.

Unfortunately, the fix in 0.14.0 has introduced another bug. Using some of the examples above, when a column of type list is shifted using `over`, we now get a list...

For documentation: Stack Overflow updated to indicate that this is resolved.

This [section](https://pola-rs.github.io/polars-book/user-guide/dsl/custom_functions.html#to-apply) from the User Guide might help: > apply works on the smallest logical elements for that operation. That is: select context -> single elements groupby context -> single...

If anyone needs temporary workarounds for set operations, I've provided them on [Stack Overflow](https://stackoverflow.com/questions/72871905/polars-intersection-of-list-columns-in-dataframe). Intersection, union, set difference, and symmetric difference. The workarounds do not involve joins, explodes, groupby, or...

> Is there any way to extend the solution to n ∩ m sets? It would be helpful if you could provide a small working example with sets of different...

On [Stack Overflow](https://stackoverflow.com/questions/73252912/how-to-set-masked-values-within-each-group-in-groupby-context-using-py-polars/73430148#73430148), I've prototyped an approach (with working code) that will hopefully strike an acceptable balance among: - ease-of-use (including convenience methods) - impact to code (very little) -...

FYI, I was just doing the first one, to finally transfer a nearly 6 billion row dataset currently stored in an IPC file to a parquet file. ```python3 import polars...

Unfortunately, we still have a problem in 0.14.1. If we start with this data: ```python3 import polars as pl df = pl.DataFrame( { "group": ["a"] * 3 + ["b"] *...