BUG: df.join(df2, how='right') TypeError #11545
Conversation
jreback
added
Bug
Indexing
| self.assertTrue(res.equals(eres)) | ||
| tm.assert_numpy_array_equal(lidx, elidx) | ||
| tm.assert_numpy_array_equal(ridx, eridx) | ||
|
|
||
| """ |
There was a problem hiding this comment.
what is still commented out here?
There was a problem hiding this comment.
I don't know. They are the remnants of some old tests that had been commented because they were not working (at least, it seems like it). I've corrected and adapted the most part, but I have no idea what the rest was intended for.
I've had a conservative approach, and kept them, but sure I can do a bit of cleaning and remove them. What should I do?
There was a problem hiding this comment.
ok, see what they do (and of course if they make sense). I suspect that the code is actually working, so adjust the tests if you would.
|
Ok, the commented test is redundant with what is just before, so I'm removing it. |
|
pls add a whatsnew note (bug fixes) & squash |
|
Done |
|
can you add all of the examples from the issue as well. (in |
|
I've put them in what I thought was the most appropriate test method, I hope that's all right. |
|
merged via d78266e thanks! |
|
Thanks for closing this out! 👍 |
Issue #11519
Somehow right joins had been forgotten in a previous bugfix. There were tests already written that should have seen the problem, but they had been commented out because the expected results were wrong, because of a subtlety in the way non-unique index are handled.
Thanks for the labels "difficulty novice" and "effort low": I'm using pandas every day, and I'm glad I could contribute that easily.