ENH: Add `complement` argument to `(Series|DataFrame).filter` to do reversely operation

[Zeroto521]

Feature Type

• Adding new functionality to pandas

• Changing existing functionality in pandas

• Removing existing functionality in pandas

Problem Description

Assume there is a DataFrame and has a lot of columns.
I want to drop some of the columns.

>>> df = pd.DataFrame(np.array(([1, 2, 3], [4, 5, 6])),
...                   index=['mouse', 'rabbit'],
...                   columns=['one', 'two', 'three'])
>>> df
        one  two  three
mouse     1    2      3
rabbit    4    5      6

# Want to drop the columns containing 'o'

>>> df.filter(like='o', complement=True)
        three
mouse       3
rabbit      6

Feature Description

def filter(
    self: NDFrameT,
    items=None,
    like: str | None = None,
    regex: str | None = None,
    axis=None,
    complement: bool = False,
):
    nkw = com.count_not_none(items, like, regex)
    if nkw > 1:
        raise TypeError(
            "Keyword arguments `items`, `like`, or `regex` "
            "are mutually exclusive"
        )

    if axis is None:
        axis = self._info_axis_name
    labels = self._get_axis(axis)

    if items is not None:

        def f(x) -> bool_t:
            return (x not in labels) if complement else (x in labels)

        name = self._get_axis_name(axis)
        return self.reindex(**{name: [r for r in items if f(r)]})
    elif like:

        def f(x) -> bool_t:
            assert like is not None  # needed for mypy
            return like in ensure_str(x)
    elif regex:

        matcher = re.compile(regex)
        def f(x) -> bool_t:
            return matcher.search(ensure_str(x)) is not None
    else:
        raise TypeError("Must pass either `items`, `like`, or `regex`")

    values = labels.map(f)
    if complement:
        return self.loc(axis=axis)[~values]
    return self.loc(axis=axis)[values]

Alternative Solutions

• For DataFrame.drop, it requires the full names of columns.

The following solutions could work but they may be a little bit annoying.

• Directly use not in, like df.loc[:, df.columns.map(lambda x: 'xxx' not in x)].
• It also could use re to search these columns and then drop them.

Additional Context

No response

[jreback]

-1 on changing filter (which we should deprecate anyhow)

[mroeschke]

Thanks for the request but it appears there hasn't been much interest in this issue from the core team and community over the years so closing.