filter

skrub.selectors.filter(predicate, *args, **kwargs)

Select columns for which predicate returns True.

For each column col in the dataframe, predicate is called as predicate(col, *args, **kwargs) and the column is kept if it returns True. To filter columns based only on their name, see also filter_names.

args and kwargs are extra parameters for the predicate. Storing parameters like this rather than in a closure can help using an importable function as the predicate rather than a local one, which is necessary to pickle the selector. (An alternative is to use functools.partial).

Examples

>>> from skrub import selectors as s
>>> import pandas as pd
>>> df = pd.DataFrame(
...     {
...         "height_mm": [297.0, 420.0],
...         "width_mm": [210.0, 297.0],
...         "kind": ["A4", "A3"],
...         "ID": [4, 3],
...     }
... )
>>> df
   height_mm  width_mm kind  ID
0      297.0     210.0   A4   4
1      420.0     297.0   A3   3

>>> selector = s.filter(lambda col: 'A4' in col.values)
>>> s.select(df, selector)
  kind
0   A4
1   A3

>>> def contains(col, value):
...    return value in col.values

>>> selector = s.filter(contains, 3)
>>> selector
filter(contains, 3)

>>> s.select(df, selector)
   ID
0   4
1   3