Selector#

class skrub.selectors.Selector[source]#

Pattern for matching columns in a dataframe.

A Selector is a reusable rule for selecting columns based on various criteria (data type, name pattern, content properties, etc.). Selectors enable delayed selection: you can define a selection rule before the data is available.

How to use selectors

Direct selection: s.select(df, selector) returns a filtered dataframe
With skrub.ApplyToCols: ApplyToCols(transformer, cols=selector) applies a transformer to selected columns
In DataOps: skrub.X(df).skb.apply(transformer, cols=selector)
Manual expansion: selector.expand(df) gets column names for manual use

Combining Selectors

Selectors can be combined with set operators to create complex selection rules:

s.numeric() | s.boolean() - numeric OR boolean columns
s.all() - s.glob('*_id') - all columns except those whose name ends with “_id”
~s.cardinality_below(10) - high-cardinality columns

Note

This class is not meant to be instantiated manually. Create selectors using builder functions such as skrub.selectors.all(), skrub.selectors.cols(), skrub.selectors.glob(), etc.

See also

skrub.ApplyToCols: Apply a transformer only to some columns, possibly selected by a selector.
skrub.DataOp.skb.apply: Apply a transformer to selected columns in a DataOps workflow.

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],
... })

Use a selector to get the names of matching columns:

>>> s.numeric().expand(df)
['height_mm', 'width_mm', 'ID']

Selectors can be combined to create complex selection rules. For example, to select numeric columns but exclude the ‘ID’ column:

>>> (s.numeric() - 'ID').expand(df)
['height_mm', 'width_mm']

Use in data transformations, for example to apply a scaler only to numeric columns:

>>> from skrub import ApplyToCols
>>> from sklearn.preprocessing import StandardScaler
>>> ApplyToCols(StandardScaler(), cols=s.numeric()).fit_transform(df)
kind  height_mm  width_mm   ID
0   A4       -1.0      -1.0  1.0
1   A3        1.0       1.0 -1.0

Methods

`expand`(df)	Get the names of columns matched by the selector in a list.
`expand_index`(df)	Get the indices of columns matched by the selector in a list.

expand(df)[source]#

Get the names of columns matched by the selector in a list.

This method evaluates the selector’s matching criteria against each column in the dataframe and returns the names of columns that match. This can be useful to extract column names to be used in dataframe operations.

Parameters:

dfdataframe: A pandas or polars dataframe to evaluate the selector against.

Returns:

list of str: The names of columns from df that the selector matches.

See also

expand_index: Get indices of matching columns instead of names.

Examples

>>> import pandas as pd
>>> from skrub import selectors as s
>>> some_selector = ~s.glob("*_mm")
>>> df = pd.DataFrame(
...     {
...         "height_mm": [210.0, 297.0],
...         "width_mm": [188.5, 210.0],
...         "kind": ["A5", "A4"],
...         "ID": [5, 4],
...     }
... )
>>> some_selector.expand(df)
['kind', 'ID']

Use to select columns in a dataframe:

>>> df[some_selector.expand(df)]
   kind  ID
0   A5   5
1   A4   4

expand_index(df)[source]#

Get the indices of columns matched by the selector in a list.

This method evaluates the selector against each column and returns the positional indices (0, 1, 2, …) of matching columns instead of their names.

Parameters:

dfdataframe: A pandas or polars dataframe to evaluate the selector against.

Returns:

list of int: The indices (zero-indexed) of columns from df that the selector matches.