.. _auto_dataset_get_group_parts_table:

*RE-ORGANIZE FEATURE IMPORTANCE LIKE TABLE TO GROUP-PARTS TABLE*
================================================================

Reorganizes a feature-importance-style flat table into a compact Group-Parts Table by grouping rows on a chosen group column, sorting features within each group by descending value, and serializing them as colon-delimited part strings. Optionally normalizes each feature's value as a ratio of the group's total. Use this worker when you need a human-readable or downstream-consumable summary of which features contribute most within each group interval.

When to use
-----------

Tagged: ``feature-importance``, ``group-parts``, ``interval``, ``pivot``, ``ratio``, ``reshape``, ``rollup``.

Inputs
------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20 20 20

   * - Label
     - ID
     - Type
     - Default
     - Required
     - Description
   * - Dataset
     - dataset
     - dataset
     - —
     - ✓
     - Input dataset in tabular (array-of-rows) format containing at minimum the group, feature, and value columns; must be a non-empty dataset object.
   * - Group Column
     - group_column
     - text
     - —
     - 
     - Name of the column whose distinct values define the groups (e.g. an interval or category ID); select a text or numeric column from the linked dataset.
   * - Feature Column
     - feature_column
     - text
     - —
     - 
     - Name of the column that identifies each feature or part within a group (e.g. component name or variable label); select a text or numeric column from the linked dataset.
   * - Value Column
     - value_column
     - text
     - —
     - 
     - Name of the column holding the numeric importance or contribution value for each feature row; select a text or numeric column from the linked dataset.
   * - Return Ratio On Each Interval
     - return_ratio
     - text
     - —
     - 
     - Boolean flag (default true); when true each feature's value is normalized as a fraction of the group's total sum and the output encodes 'feature:ratio=X', otherwise raw values are used with 'feature:value=X'.

Outputs
-------

.. list-table::
   :header-rows: 1
   :widths: 20 20 20 20

   * - Label
     - ID
     - Type
     - Description
   * - Output Group Parts Table
     - dataset
     - dataset
     - Output Group-Parts Table: one row per group containing the group-column value and a 'Parts' string of '::'-delimited 'feature:ratio=X' (or 'feature:value=X') entries sorted by descending importance.

Disciplines
-----------

- ai_ml.prognosis
- data.dataset.transform
- data.statistics

Runnable example
----------------

A runnable example is registered for this worker. Open the example workflow on the d3VIEW canvas: `/api/workflow/example?id=dataset_get_group_parts_table <https://www.d3view.com/api/workflow/example?id=dataset_get_group_parts_table>`_

.. raw:: html

   <hr style="margin-top:2em">
   <p style="font-size:11px;color:#888">
   Auto-generated from <code>platform</code> schema. Worker id: <code>dataset_get_group_parts_table</code>. Schema hash: <code>840b3b15a7d4</code>. Hand-curated docs in <code>workerexamples/</code> override this page when present.
   </p>