Tabular: pivot collection to wide

Tool

toolshed.g2.bx.psu.edu/repos/iuc/collection_column_join/collection_column_join/0.0.3. The tabular survey identifies this as the dominant IWC wide-pivot idiom: a collection of per-element (identifier, value) tables becomes one wide table with one row per identifier and one column per collection element.

When to reach for it

Use this when each collection element is a two-column tabular keyed by the same identifier column, and the desired output is one wide tabular matrix.

Do not use this for ordinary two-file joins; use tabular-join-on-key. Do not frame it as a generic transpose or pivot-table operation; the survey found no broad datamash_transpose-style pattern in IWC.

Parameters

• input_tabular: connected collection of tabular datasets.
• identifier_column: 1-indexed key column in each per-element table. Corpus examples usually use "1".
• fill_char: value for missing element/key cells. Corpus examples use "0" and ..
• has_header: string toggle, "0" / "1", not a boolean.
• old_col_in_header: whether collection element / original column identity appears in output headers.
• include_outputs: usually null in corpus examples.

Idiomatic shapes

Headerless abundance/count pivot:

tool_id: toolshed.g2.bx.psu.edu/repos/iuc/collection_column_join/collection_column_join/0.0.3
tool_state:
  fill_char: "0"
  has_header: "0"
  identifier_column: "1"
  include_outputs: null
  input_tabular: { __class__: ConnectedValue }
  old_col_in_header: true

Anchored by the MAPseq-to-ampvis2 IWC exemplar.

Headered metric-table pivot:

tool_id: toolshed.g2.bx.psu.edu/repos/iuc/collection_column_join/collection_column_join/0.0.3
tool_state:
  fill_char: .
  has_header: "1"
  identifier_column: "1"
  include_outputs: null
  input_tabular: { __class__: ConnectedValue }
  old_col_in_header: false

Anchored by the MAGs generation IWC exemplar.

Pitfalls

• Input shape is strict. This works because upstream emits one (id, value) table per collection element. It is not a generic wide/long pivot.
• Do not substitute ordinary joins. collection_column_join is collection-shaped; use tabular-join-on-key for two-file key joins.
• has_header is a string. Corpus YAML uses "0" / "1", not booleans.
• Fill value has meaning. "0" means absent is zero; . means missing/unknown. Match downstream semantics.
• Header semantics matter. old_col_in_header: true preserves element identity in headers; false appears in MAG metric pivots.
• Guard empties when plausible. The transformations survey flags upstream __FILTER_EMPTY_DATASETS__ as a defensive guard when per-element outputs may be empty, especially small-N or filter-heavy paths. Do not apply it as universal boilerplate; several IWC pivots do not filter first.

Legacy alternative

None. This page exists because the corpus has one dominant operation-shaped path. Awk or SQL rewrites would be custom transformations, not the IWC pattern.

See also

• iwc-tabular-operations-survey — candidate 8 and §7 operation-anchored naming decision.
• iwc-transformations-survey — collection-side cross-reference and empty-element guard note.
• tabular-join-on-key — ordinary two-file key joins.
• tabular-concatenate-collection-to-table — row-bind a collection into one long table.