Configuration

Splitting & Joining Data

Learn how to split documents into pages and join results back together.

Updated 12/11/2025

bem provides primitives for handling lists and aggregations through Split and Join functions.

Split Function

A Split Function (type: "split") breaks down an input (like a document or a list) into smaller chunks for processing.

Split Types

  • print_page: Splits a document by physical pages.
  • semantic_page: Splits based on semantic items defined in semanticPageSplitConfig.

Join Function

A Join Function (type: "join") aggregates results from previous steps.

Configuration

  • joinType: Currently supports standard.
  • outputSchema: Defines the structure of the joined output.

These primitives are essential for "Fan-out/Fan-in" workflow patterns.