Empty KeyError when trying to merge Int32 and int32[pyarrow] columns

[canutog]

Hi, I'm using Pandas 2.0.0. When I tried to merge two int columns('int32', 'int32[pyarrow]') I recieved an empty KeyError.

When I try to merge columns of different dtypes like int and object the KeyError has an explicit message saying that you're trying to merge with two columns of different types.

But this does not happen for int32 and int32[pyarrow] or float for similar types between old format and pyarrow new format.

Could you please update, so that this error can be identified more easily? I spent aproximately 4h figuring this out.
Thanks

code_to_reproduce_error.txt

[phofl]

Hi, thanks for your report. Please provide a minimal and reproducible example, see https://matthewrocklin.com/minimal-bug-reports

Your example is too large. Also please fill out our issue template in the future.

[canutog]

Hi, my code is in the txt file in the issue page. Em ter., 4 de abr. de 2023 14:24, Patrick Hoefler ***@***.***> escreveu:

…

Hi, thanks for your report. Please provide a minimal and reproducible example, see https://matthewrocklin.com/minimal-bug-reports Your example is too large. Also please fill out our issue template in the future. — Reply to this email directly, view it on GitHub <#52406 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQF4NFKJFPH3WGBWYE6ZI3TW7RKNZANCNFSM6AAAAAAWS2EBYE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[phofl]

This is not a minimal example, this is way too much code

[canutog]

In pandas 1.5.3, this example works fine. With Pandas 2.0.0, the example returns a key error without any information
error_example.txt