base.group_by

Group a table by one or more columns and compute aggregations per group. This is a convenience wrapper around base.aggregate that promotes the grouping columns to a dedicated parameter so pipelines read more naturally.

Kind: standard
Network: no

Inputs🔗

Name	Type	Description
`table`	file (Parquet, CSV accepted)	Source table.
`group_columns`	string	Comma-separated list of grouping columns (required, non-empty).
`aggregations`	string	JSON list of aggregation specs (same format as `base.aggregate`).

Outputs🔗

Name	Type	Description
`result`	file (Parquet)	Grouped table with one row per group.

Example🔗

- id: 01890000-0000-0000-0000-000000000004
  name: base.group_by
  inputs:
    - 01890000-0000-0000-0000-000000000000
  args:
    group_columns: "state,county"
    aggregations: '[{"column": "population", "function": "sum", "as": "pop"}]'