BUG: MultiIndex slicing with negative step by lukemanley · Pull Request #46156 · pandas-dev/pandas (original) (raw)

• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added an entry in the latest doc/source/whatsnew/v1.5.0.rst file if fixing a bug or adding a new feature.

MultiIndex label-based (.loc) slicing with a negative step size appears to contain a few bugs.

#38071 may have partially addressed negative step size issues, but a few issues seem to remain.

I believe this PR fixes the following:

Bug 1: slicing with non-null start/stop values returns an empty frame

    from pandas import DataFrame, IndexSlice, MultiIndex

    mi = MultiIndex.from_arrays([range(5)] * 2)
    df = DataFrame(1.0, index=mi, columns=['A'])

    target = IndexSlice[3:1:-1, :]
    df.loc[target, :]

Bug 2: negative step size other than -1 raises

    from pandas import DataFrame, IndexSlice, MultiIndex

    mi = MultiIndex.from_arrays([range(5)] * 2)
    df = DataFrame(1.0, index=mi, columns=['A'])

    target = IndexSlice[::-2, :]
    df.loc[target, :]

Bug 3: slicing a non-int labeled index level raises

    from pandas import DataFrame, IndexSlice, MultiIndex

    mi = MultiIndex.from_arrays([list('abcde')] * 2)
    df = DataFrame(1.0, index=mi, columns=['A'])

    target = IndexSlice['d':'b':-1, :]
    df.loc[target, :]

The tests have been expanded to cover these cases.