Purpose
Sequentially combines rows from 2 to 32 sources. By name, it builds a shared structure from all inputs; by position, it aligns fields by their location. It does not look for matching keys or add another table’s attributes to an individual row.
Inputs and outputs
A separate named input for each table; one out output. All inputs must be connected, even when some branches are inactive during a run.
Parameters
| Parameter | Settings and values | Options and limits |
|---|---|---|
inputs | Named inputs, each receiving a separate connection. | Count: 2–32 |
mode | by_name aligns matching names and fills missing fields with NULL; position aligns columns by position and requires the same column count. | Options: by_name, positionDefault: by_name |
distinct | Remove entirely identical result rows, taking types into account. | Default: disabled |
Step-by-step configuration
- Create inputs and give them clear names.
- Connect each source to a separate input.
- Choose alignment by name or position. For position, check each input’s own field order.
- Enable distinct if needed, then review the combined headings and row count.
Rules and diagnostics
- Input order determines the order in which rows are appended.
- By name, fields from the first input come first, followed by new fields from later inputs.
- By position, unequal column counts cause UNION_SCHEMA. Result headings come from the first input.
- distinct compares the entire row. To remove duplicates using only some fields, use “Deduplication”.
Configuration example
Input 1 has Product, Value; input 2 has Product, Value, Region. by_name produces Product, Value, Region; Region is NULL for rows from the first input.
Per-input field order
Each input can define order, a field list in the desired sequence. Listed fields come first, followed by the remaining fields in their original order. Listed fields absent from the current table are skipped. In position mode, check every input’s order first: an equal column count does not guarantee equal meaning.