Join two tables on one or more shared key columns. The binary counterpart to base.reduce_join (which gathers a map fan-in collection): use base.join to merge two independently-produced tables, e.g. a table of area estimates with a table of covariates. Overlapping non-key columns from the right table are suffixed with _right. Accepts CSV or Parquet input and emits Parquet output.
Right table. Because both inputs share the same type, a pipeline must wire them with explicit references that name the target input via as (e.g. as: left / as: right), not bare ones.
on
string
Comma-separated key column name(s) that must exist in both tables.
how
string
Join type - one of "inner" (default), "left", "right", or "outer".