The Union node stacks rows from two data streams vertically into a single output — like appending one spreadsheet below another. Unlike Merge (which joins data side-by-side), Union appends data vertically. It can optionally deduplicate records based on mapped fields.
Basic Example
Input 1
Input 2
Result (Union)
How it works
Drag two data streams into the Union node. These can come from any upstream nodes in your workflow. Choose a union type
Select how the rows from both inputs should be combined.| Union Type | SQL Equivalent | Description |
|---|
| All | UNION ALL | Keep all rows from both inputs — simple concatenation (default) |
| Distinct | UNION | Deduplicate rows based on the mapped fields, keeping one row per unique combination |
When using Distinct mode, configure which fields define uniqueness and which input wins on duplicates.Dedup Key — select which fields from each input correspond to each other. These field pairs define uniqueness for deduplication. You can add multiple fields for a composite key (e.g., ID + Date).| Setting | Description |
|---|
| Input 1 Field | A field from the first data stream |
| Input 2 Field | The corresponding field from the second data stream |
| Priority | Description |
|---|
| Input 1 wins | Keep the row from Input 1 when a duplicate is found (default) |
| Input 2 wins | Keep the row from Input 2 when a duplicate is found |
Field mappings are required when using Distinct mode. In All mode they are optional — rows are simply stacked.
Output
The Union node produces a single output stream containing the combined rows from both inputs. The output is available as an input to any downstream node in your workflow.
Example with Deduplication
Input 1
Input 2
Configuration: Distinct, Dedup Key: ID, Priority: Input 1 wins
Result:
All vs Distinct
When using All, if both tables contain the same record, the result will include it twice:
When using Distinct, duplicates are removed based on the dedup key, keeping only one row per unique combination.
Union vs Merge vs Bond
| Feature | Bond | Merge | Union |
|---|
| Purpose | Relate data | Combine columns | Stack rows |
| Structure | Linked | Flattened | Appended |
| SQL Equivalent | Join (logical) | Join (materialized) | UNION / UNION ALL |
| Use Case | Enrichment | Final dataset | Combining datasets |
Best Practices
- Ensure inputs have compatible schemas (same fields and meaning)
- Use All when you want full data coverage and duplicates are acceptable
- Use Distinct when you need a clean dataset and duplicate records must be removed
- Carefully define dedup keys and priority rules
Common Pitfalls
- Mismatched columns — same name but different meaning across inputs
- Missing dedup key — leads to incorrect duplicate removal
- Over-aggressive deduplication — can cause data loss
- Merge Node — joins two datasets side-by-side rather than stacking vertically
- Bond Node — creates logical relationships between entities
