Union Transform

Union appends rows from two or more datasets, lining up their columns by name, position, or a manual mapping.

Basic Usage

Before configuring the union, connect its inputs:

caution

Union requires at least two unique parent nodes connected directly to it. They can be Data Input nodes or outputs from other transforms, but at least one configured Data Input must exist upstream.

1. Select Union from the transform menu.
2. Connect the parent datasets to append.
3. Choose an Alignment Mode and set the other options.
4. If you choose Manual, configure the column mapping.

Configuration Options

Basic Options

• Alignment Mode: Choose how columns from each input line up:

  • Name: Lines up columns with the same name.
  • Position: Lines up columns by their order in each dataset.
  • Manual: Lets you map columns between datasets.

• Output Order: Drag the inputs into the order their rows should appear.

• Common Columns Only: Keep only columns present in every input.

Advanced Options

• Manual Column Mapping: In Manual mode, match columns that represent the same field.

• Column Configuration: Choose which fields to include and how to line them up.

Examples

Example 1: Name-based Alignment

Input Datasets:

Dataset 1:

A	B
1	x
2	y

Dataset 2:

B	C
z	3
w	4

Configuration:

• Alignment Mode: Name
• Common Columns Only: Off

Result:

A	B	C
1	x	null
2	y	null
null	z	3
null	w	4

Example 2: Position-based Alignment

Input Datasets:

Dataset 1:

A	B
1	x
2	y

Dataset 2:

C	D
3	z
4	w

Configuration:

• Alignment Mode: Position
• Common Columns Only: Off

Result:

A	B
1	x
2	y
3	z
4	w

Example 3: Manual Alignment

Input Datasets:

Dataset 1:

Name	Age
John	30
Mary	25

Dataset 2:

Full Name	Years
Bob Smith	40
Alice Lee	35

Configuration:

• Alignment Mode: Manual
• Manual Mapping:
  • Name → Full Name
  • Age → Years

Result:

Full Name	Years
John	30
Mary	25
Bob Smith	40
Alice Lee	35

tip

Choose Manual when differently named columns represent the same field.

caution

Position ignores column names. If the inputs order their columns differently, unrelated values can end up in the same output column.

Best Practices

1. Check how columns line up, especially in Position mode.
2. Expect nulls when an input lacks a column and Common Columns Only is off.
3. Set Output Order when row order matters.
4. Test a large union with representative samples before running the full datasets.