ScaleCategorizer

class

class ScaleCategorizer(
    column_names: ColumnNamesWithAltSpec | None = None,
    on_missing_column: OnInvalidColumnSpec = 'ignore',
    on_wrong_type: OnInvalidColumnSpec = 'ignore',
    on_wrong_value: OnInvalidValueSpec = 'ignore',
    scale: Scale = field(kw_only=True),
    tags: TagsSpec | None = field(default=None, kw_only=True),
)

Bases: Categorizer

Categories: dq-classifier

A classifier that assigns each value to a bin defined by a Scale. Applies to numeric columns. The verdict column holds the bin index (or a reserved sentinel for null / nan / under- / overflow); pair it with an InBins or NotInBins criteria.

Examples

ScaleCategorizer(["value"], scale=LinearScale(0.0, 10.0, bins=5))

Parameters

parameter

column_namesColumnNamesWithAltSpec | None (list[str | tuple[str, str | None]] | None)

Columns to classify. Each entry is a column name, or a (column, verdict_column) tuple naming the column the verdict lands in; the two forms can be mixed. None (the default) classifies every user column, skipping the types this classifier does not support (see on_wrong_type). A verdict column that is not named explicitly is the source column plus this classifier's suffix (a column amount lands in amount_category).

parameter

on_missing_columnOnInvalidColumnSpec (Literal['ignore', 'fail'])

What to do when a named column is not in the table: "ignore" (the default) leaves it out, "fail" aborts the check.

parameter

on_wrong_typeOnInvalidColumnSpec (Literal['ignore', 'fail'])

What to do when a column's data type is not one this classifier supports: "ignore" (the default) leaves it out, "fail" aborts the check.

parameter

on_wrong_valueOnInvalidValueSpec (Literal['ignore', 'fail'])

What to do when a value cannot be evaluated against this classifier: "ignore" (the default) yields a null verdict, "fail" aborts the check.

parameter

scaleScale

The Scale (e.g. LinearScale, LogarithmicScale) that maps values to bins.

parameter

tagsTagsSpec | None (list[str] | None)

Optional labels for this classifier. An operator given the same tag applies only to the classifiers carrying it; untagged operators apply to all of them.

Methods

method

supported_dtypes

def supported_dtypes() -> frozenset