base.reduce_join

Join a collection of tables on one or more key columns (the SQL JOIN analogue). Tables are joined sequentially in the collection's lexicographic filename order, accumulating left-to-right.

  • Kind: reduce
  • Network: no

Inputs🔗

NameTypeDescription
tablescollection of file (Parquet)Tables to join, processed in lexicographic order of filename.
onstringComma-separated list of key column names.
howstringOne of inner (default), left, right, outer.

Outputs🔗

NameTypeDescription
resultfile (Parquet)Joined table.

Example🔗

- id: 01890000-0000-0000-0000-00000000000c
  name: base.reduce_join
  inputs:
    - 01890000-0000-0000-0000-000000000007
  args:
    on: "id"
    how: "inner"