Clean

Bases: Transformer

              flowchart TD
              matchlab.transformers.clean.Clean[Clean]
              matchlab.transformers.base.Transformer[Transformer]

                              matchlab.transformers.base.Transformer --> matchlab.transformers.clean.Clean
                


              click matchlab.transformers.clean.Clean href "" "matchlab.transformers.clean.Clean"
              click matchlab.transformers.base.Transformer href "" "matchlab.transformers.base.Transformer"
            

Derive columns with DuckDB SQL, without dropping unrelated fields.

Each entry maps an output column name to a SQL expression over the record step's columns (the source-qualified names, crn_company). An alias that names an existing column replaces it. A new alias is added. Every other column passes through untouched. Dropping is Select's job, not Clean's.

id is not writable. It is the grouping a model matches on, so an alias naming it is refused rather than allowed to displace it.

Methods:

• apply –

  Add or replace the named columns, keeping the rest.

Attributes:

• version (int) –
• cleaning (dict[str, SQLExpression]) –
• model_config –

version: int = 1

cleaning: dict[str, SQLExpression] = Field(description='Output column name to a DuckDB SQL expression over the step.')

model_config = ConfigDict(extra='forbid', frozen=True)

apply(data: DataFrame) -> DataFrame