ENH: Add index to output of assert_series_equal on category and datetime values #33575
Conversation
| elif is_extension_array_dtype(left.dtype) and is_extension_array_dtype(right.dtype): | ||
| assert_extension_array_equal(left._values, right._values) | ||
| assert_extension_array_equal( | ||
| left._values, right._values, index_values=np.asarray(left.index) |
There was a problem hiding this comment.
Do you have to use asarray here or could you just pass back the DatetimeArray? Would make for nicer output if the date time values were shown
There was a problem hiding this comment.
That is a good point which I hadn't thought of. However, it isn't quite as simple as that because we only know that the data is an extension array at this point - the index may not be. Also, my understanding of extension array is very poor, but I thought it could be used for things other than datetimes?
The main reason though, is that the data is currently dislayed as numeric timestamps, so it seems reasonable to fix the index display at the same time as the data. And that sounds like a seperate PR to me. I am looking at doing it (when I get time), but as mentioned above, I need to figure out extension arrays first.
There was a problem hiding this comment.
And thanks for looking at the PR. Still not sure if I'm supposed to click 'Resolve conversation' now.
|
thanks @amilbourne |
3 participants
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffThis is really an extension of #31435, which added index output to the messages produced when series (and DataFrames) are compared. I failed to notice that categorical values and datetime values were evaluated through a different code path and didn't add the index output for these data types. Since the original problem (index reordering) could just as easily happen for these types the functionality should be there as well. Also, it makes the output more consistent.
I added a couple of new tests to surface and check the additional output.