Purpose

A subflow centralizes shared cleaning, calculations or mappings. Unlike a separate flow run, it executes inside the current execution and returns a table without creating a new background job.

Input and output

The block has one in input and out output. Its table goes to the selected “Receive data from flow” block in the child’s Production version. Data reaching the selected “Return data to flow” block becomes the current block’s output. Only the graph portion required by that output is evaluated.

Subflow contract

  • flow_id — a flow with published Production.
  • input_node — the child’s declared input.
  • output_node — the child’s declared output.
  • parameters — constants, current launch parameters or first-row input values.
  • The eye icon opens the child’s Production version for inspection.

Reuse setup

  1. Create a child: receive → transform → return data.
  2. Publish it and verify execution rights.
  3. Add “Call flow” in the parent and select flow, input and output.
  4. Match field names and required parameters.
  5. Preview the block output and simulate the parent process.

Subflow boundaries

The selected output’s dependency path may contain transformations, parameters, inputs, outputs and nested subflows. It performs no external reads or writes, asks for no user action and starts no other runs; use “Call flows” for such processes. Maximum depth is 5 flows, with no cycles. Preparation freezes Production and its revision; the user must be authorized for every dependent flow. Missing inputs or outputs require the reference to be corrected.

Shared-rules example

Put product-code normalization and net-amount calculation in a subflow. Excel, SQL and API processes use it after reading their own sources. After changing a rule, test the subflow, publish Production and verify dependent scenarios; each receives a consistent table.