RenamePlan#

class scikitplot.feature_selection.RenamePlan(pairs)[source]#

A complete, frozen old -> new name mapping.

Parameters:

pairstuple of (str, str): One pair per input column, in column order.

Parameters:

pairs (tuple[tuple[str, str], ...])

See also

FeatureRenamer: Applies a plan inside a pipeline.
FeatureSchema.rename: Renames a fitted schema, keeping aliases.

Examples

>>> from scikitplot.feature_selection import RenamePlan
>>> RenamePlan.build(["a", "b", "c"], "col-{index}", start=1).names_out
['col-1', 'col-2', 'col-3']
>>> RenamePlan.build(["a", "b"], {"a": "alpha"}).names_out
['alpha', 'b']
>>> RenamePlan.build(["x", "y"], str.upper).mapping
{'x': 'X', 'y': 'Y'}

apply(X)[source]#

Rename the columns of X, which must equal names_in in order.

Parameters:

Xpandas.DataFrame: Frame to rename.

Returns:

pandas.DataFrame: A new frame with renamed columns (data not copied where pandas avoids it).

Raises:

ValueError: If the columns differ from the plan’s inputs.

Parameters:

X (DataFrame)

Return type:

DataFrame

classmethod build(names, rule, *, start=0, on_collision='raise', on_missing='raise')[source]#

Apply rule to names once and freeze the result.

Parameters:

namessequence of str: Input column names (unique).
rulemapping, sequence of str, callable, str or None: See the module documentation. None keeps every name.
startint, default=0: First {index} of a template.
on_collision{“raise”, “suffix”}, default=”raise”: Handling of duplicate new names.
on_missing{“raise”, “ignore”}, default=”raise”: Handling of mapping keys that are not in names.

Returns:

RenamePlan: The plan.

Raises:

ValueError: For a list of the wrong length, an unknown template field, an unknown mapping key, an empty new name or a refused collision.
TypeError: For a rule of an unsupported type or a callable returning a non-string.

Parameters: