All Products
Search
Document Center

Platform For AI:Merge rows (UNION)

Last Updated:Apr 01, 2026

The Merge Rows (UNION) component stacks two tables vertically into a single output table. It supports both UNION (removes duplicate rows) and UNION ALL (keeps all rows, including duplicates), controlled by the Remove Duplicates parameter.

Both input tables must have the same number of columns and matching data types.

Note

Configure this component only in the Machine Learning Platform for AI (PAI) console.

Configure the component

In Machine Learning Designer, open your pipeline and add a Merge Rows (UNION) component. Connect it to the upstream nodes, then click the component to configure the following parameters.

ParameterDescription
Left Table Output ColumnColumns to include from the left input table. The number of selected columns and their data types must match the right table selection.
Right Table Output ColumnColumns to include from the right input table. Must match the left table selection in count and data type.
WHERE Clause for Left TableFilters rows from the left table before merging. Uses SQL-style syntax. Example: age>40.
WHERE Clause for Right TableFilters rows from the right table before merging. Uses the same syntax as the left table WHERE clause.
Remove DuplicatesControls whether duplicate rows are removed from the output. Selected by default (equivalent to UNION). Clear this check box to keep all rows, including duplicates (equivalent to UNION ALL).

WHERE clause operators

The WHERE clause supports the following comparison operators:

OperatorExample
=age=40
!=age!=40
>age>40
<age<40
>=age>=40
<=age<=40
likename like 'Ali%'
rlikename rlike '^Ali.*'