Group

Bases: Transformer

              flowchart TD
              matchlab.transformers.group.Group[Group]
              matchlab.transformers.base.Transformer[Transformer]

                              matchlab.transformers.base.Transformer --> matchlab.transformers.group.Group
                


              click matchlab.transformers.group.Group href "" "matchlab.transformers.group.Group"
              click matchlab.transformers.base.Transformer href "" "matchlab.transformers.base.Transformer"
            

Collapse each id to a single row, combining columns with aggregate SQL.

Reading several records per entity (through a resolver) gives a model more evidence, but a comparison needs one row. Group says how each column combines. id is the grouping key, and every named expression is a DuckDB aggregate, for example any_value(crn_company) where records agree or list(distinct crn_town) where they don't. Only id and the named aggregates survive, since grouping has to decide how every column collapses.

Methods:

• apply –

  Collapse each id to one row using the aggregate expressions.

Attributes:

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

version: int = 1

aggregates: dict[str, SQLExpression] = Field(description='Output column name to a DuckDB aggregate expression.')

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

apply(data: DataFrame) -> DataFrame