reswqa
reswqa
That's great and exactly what I'm looking forward to. Before seeing this, I used to rewrite Python code into equivalent Rust code to debug.(*´I`*)
@mcrumiller I'm a jetBrains user, I can help verify it, but I suggest not blocking this PR. :)
Thanks for reporting this! I believe the problem comes from `gather_every` itself rather than `list.eval`.
> Is there something else required to get this merged? :) No, but we need to wait for Ritchie to make the final review.
I am sorry that an unrelated PR was accidentally linked to this issue, please ignore the above message(mentioned this issue). :)
I'm not sure if we can implement a good hash function for anonymous functions, maybe @ritchie46 has some ideas. But as for this issue itself, I think we can make...
> Is it correct that that PR solves this issue specifically for duration but not for other anonymous functions? Yes, this is why I marked this PR as `refactor` instead...
@mcrumiller Great, it works without raise. Thank you for noticing this!
> But I prefer what we do at str.concat. That is, we introduce the ignore_null argument to decide whether to ignore or propagate null values. This shouldn't block the PR....
They have a clear distinction mainly in the `agg` context. If `agg_list` is `False`, the UDF is called per group. In contrast, the UDF is invoked only once on a...