BUG: Join behaved like an inner join when only one side had MultiIndex

[phofl]

• closes BUG: left join between df with a single index and df with a multiindex produces an inner join #34292
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

We can use the regular path now, since droplevel can handle an Index. Only difference is a resorted result, because Index.join() does not preserve the order of the leading input

left = pd.Index([0, 1, 2, 1], name="b")
right = pd.Index([1, 2, 5], name="b")
print(left.join(right, how="left"))

returns

Int64Index([0, 1, 1, 2], dtype='int64', name='b')

Edit: We may be able to delete the _join_level function, if we add a level keyword to the _join_multi function to determine the overlap.

[phofl]

The change caused a few other tests to fail, which I have not run locally. All because the new result has a differently sorted index. Is this ok or do we have to figure a way out to get the old behavior?

[github-actions]

This pull request is stale because it has been open for thirty days with no activity. Please update or respond to this comment if you're still interested in working on this.

[jreback]

@phofl if you can merge master will have a look

[jreback]

move to 1.3

[phofl]

Done

[jreback]

@phofl status here?

[simonjayhawkins]

@phofl status here?

[simonjayhawkins]

@phofl closing as stale. reopen when ready.