lale.lib.rasl.aggregate module

class lale.lib.rasl.aggregate.Aggregate(*args, _lale_trained=False, _lale_impl=None, **kwargs)

Bases: TrainedIndividualOp

Relational algebra aggregate operator.

This documentation is auto-generated from JSON schemas.

Parameters

• columns (union type, optional, not for optimizer, default []) –

  Aggregations for producing output columns.

  • dict

    Dictionary of output column names and aggregation expressions.

  • or array of items : expression

    List of aggregation expressions. The output column name is determined by a heuristic based on the input column name and the transformation function.

• group_by (union type, optional, not for optimizer, default []) –

  Group by columns for aggregates.

  • expression

    Expressions for columns name if there is a single column.

  • or array of items : expression

    List of expressions for columns.

• exclude_value (Any, optional, not for optimizer, default None) – Exclude this value in computation of aggregates. Useful for missing value imputation.

transform(X, y=None)

Transform the data.

Parameters

X (array, >=1 items of items : array of items : Any) – Output of the group by operator - Pandas / Pyspark grouped dataframe

Returns

result – The outer array is over rows.

• items : array of items : Any

  The inner array is over columns.

Return type

array