base.pivot

Reshape a table between long and wide form. In "wider" mode the distinct values of names_from become new columns, filled from values_from (one value per id/name cell). In "longer" mode the values_from columns are gathered into a variable/value pair keyed by id_columns. Accepts CSV or Parquet input and emits Parquet output.

Kind: standard
Network: no

Inputs🔗

Name	Type	Description
`table`	file (Parquet, CSV accepted)	Source table.
`direction`	string	`"wider"` (default) or `"longer"`.
`id_columns`	string	Comma-separated columns that identify a row and are preserved on both sides of the reshape.
`names_from`	string	wider: the column whose distinct values become new column names. longer: unused.
`values_from`	string	wider: the column whose values fill the new columns. longer: comma-separated columns to gather (empty = all non-id columns).
`fill`	string	wider only: numeric value to substitute for empty cells (e.g. `"0"` to match the reference workflow's NA-to-zero on height bins). Empty (default) leaves them null.

Outputs🔗

Name	Type	Description
`result`	file (Parquet)	The reshaped table.

Example🔗

- id: 01890000-0000-0000-0000-00000000000e
  name: base.pivot
  inputs:
    - 01890000-0000-0000-0000-000000000000
  args:
    direction: "wider"
    id_columns: "plot_id"
    names_from: "height_bin"
    values_from: "count"
    fill: "0"