avm19

I can confirm that this is still an issue. Restarting Notepad++ does not fix this behaviour. This problem started affecting me very recently (within one month). Here's my debug info:...

@mere-human A newer version has been working well since I installed it 16 hours ago. ``` Notepad++ v8.2.1 (32-bit) Build time : Jan 19 2022 - 18:38:49 Path : C:\Program...

@vanclute As I mentioned above, I've had this problem, but with ```Notepad++ v8.2.1``` I currently experience no problems. Maybe you should try downgrading to this version? This can potentially solve...

@AdamMaras Any comment please? Would you endorse @DrMerfy's fork? Thx! https://github.com/DrMerfy/vscode-overtype

First of all, @moi90 example has a mistake, which does not invalidate it, but still. Instead of > ```python > counts.loc[(pd.NA, True, False)] # KeyError: (, True, False) > ```...

Okay, this indeed seems to be related to `GroupBy` and not just to `MultiIndex`. Here is another illustrative example: ```python df = pd.DataFrame( [[False, None, False], [True, None, False], [True,...

I think I solved it. I can open a PR soon, in a week or two, meanwhile I'll describe my findings. The problem is in `BaseMultiIndexCodesEngine.get_loc()` method": https://github.com/pandas-dev/pandas/blob/b162331554d7c7f6fd46ddde1ff3908f2dc8bcce/pandas/_libs/index.pyx#L807-L816 The mistake...

> copying my comment from #19771 ... > ... the integer coding backing a MultiIndex uses -1, so to be properly lexsorted, NA must be in the first position. Let...

I believe the code you are complaining about is in these lines (but it's worth double-checking): https://github.com/pandas-dev/pandas/blob/bef88efe999809b775ed88a02f0fc2fd6d2d08a2/pandas/core/_numba/kernels/min_max_.py#L106 https://github.com/pandas-dev/pandas/blob/bef88efe999809b775ed88a02f0fc2fd6d2d08a2/pandas/core/_numba/kernels/min_max_.py#L49-L52 ------ ~~Let me explain how I think your issue report could be...

Are there any temporary workarounds? Thanks.