Closed andresliszt closed 1 year ago
xref https://github.com/pandas-dev/pandas/issues/50564
I'm not saying tha this issue needs to fix the pandas problem I reported, but I was going to look back into it and noticed that things are in flux. When less in flux, I'll look harder.
I'm not sure why in the actions we are getting the pandas from
pandas: ["git+https://github.com/andrewgsavage/pandas.git@assert_1_5",
"git+https://github.com/andrewgsavage/pandas.git@assert_almost_equal", ]
Looking at the logs of my PR seems that it installed a dev version of pandas 0.24.0
Installing collected packages: pandas
Attempting uninstall: pandas
Found existing installation: pandas 2.0.1
Uninstalling pandas-2.0.1:
Successfully uninstalled pandas-2.0.1
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
pint-pandas 0.4.dev156+g27270a6 requires pandas>=1.5, but you have pandas 0.24.0.dev0+13114.g7380206be which is incompatible.
Successfully installed pandas-0.24.0.dev0+13114.g7380206be
I changed the pandas matrix to test on pandas >= 1.5.0
This looks good. Could you fix the conflicts and I'll merge after?
bors r+
Build succeeded!
The publicly hosted instance of bors-ng is deprecated and will go away soon.
If you want to self-host your own instance, instructions are here. For more help, visit the forum.
If you want to switch to GitHub's built-in merge queue, visit their help page.
This PR tries to support
pint 0.21
. This PR also considers a strange behavior that was not appreciated in minor pandas versions.The module
pandas.core.nanops
is supposed to act on numpy ndarrays, that's why in the testTestGroupby.test_in_numeric_groupby
we were getting the index containing the column"C"
only. Minor pandas versions were not raising errors and they were just omitting the column passed to thenanops
functions, but nowpandas >=1.5.3
is raising error. To solve this i added anumpy_data
property that tries to pass numpy ndarrays to thenanops
functions