BUG: IntervalIndex.get_indexer raising for read only array #53703
Conversation
|
Should we just generally go back to using ndarray instead of memoryviews? If we set that expectation could be a good change for new contributors to pick up |
|
This change is only necessary because cython does not support const with fused types unfortunately |
|
Ah OK. So I guess the pattern is for fused types prefer ndarray (unless that gets fixed/implemented) elsewhere memview? |
|
Yep that makes sense I guess |
|
|
||
| result = idx.get_indexer_non_unique(arr)[0] | ||
| expected = np.array([0, 1]) | ||
| tm.assert_numpy_array_equal(result, expected, check_dtype=False) |
There was a problem hiding this comment.
Hmm what's going on with the dtype? Wouldn't expect that issue
There was a problem hiding this comment.
Windows/non windows stuff and since the dtype really doesnt matter here, I just ignored it
There was a problem hiding this comment.
Maybe you can explicitly specify the type of the np.array call instead?
There was a problem hiding this comment.
This would kill the 32 bit build... Open for suggestions but dtype really doesn't matter
And that's also what we already do in most places. For example in groupby.pyx, you will see that we typically use ndarray for the input values (which use a fused type and can be const), while mostly use memoryviews for output values (which are writeable) or labels/mask (which don't use fused types). |
|
merging this, happy to address comments in a follow up. |
doc/source/whatsnew/vX.X.X.rstfile if fixing a bug or adding a new feature.