Skip to main content
The Limit transformation caps the number of output records, either for the whole input or for each group of records that share the values of selected fields.
Limit component configuration panel

Data partitioning

Select treat the entire input as a single partition to limit the output of the entire input. Select partition data by fields to limit the output per group of records in the input with identical values in selected fields.

Data order

Select don't sort input data if you’d like to output arbitrarily selected records per partition. Select sort partition by fields to select fields to sort each partition by and the direction (ascending or descending). Null values sort lowest: first in ascending order, last in descending order. So with an ascending sort and a limit of 1, a record whose sort field is null is kept ahead of the record with the lowest real value. To prefer records that have a value, see Null values in Order by fields.

Rows with empty key

When you partition data by fields, choose what happens to rows whose partition key is empty. A key is empty when any partition field is null, or is blank after trimming spaces. When you partition data by fields you can also turn on Prefer the most complete row. Each partition is then ordered first by how many of the chosen fields are filled; see Keep the most complete row.

Limit data

Finally, select the number of output records per partition.
Setting the number of output records per partition

Remove duplicates

To keep one row per key, for example the latest record for each customer, use the Dedupe transformation. It is a Limit set to 1 row per partition, with labels for that job and a default that keeps rows with an empty key.

FAQ

Q: What happens to rows with an empty partition key? By default they are grouped and limited like any other partition. Set Rows with empty key to keep all of them unchanged, or to drop them. Q: How do I keep only the latest row per key? Use the Dedupe transformation: choose the key fields, order by the date field, and keep the last row.

Dedupe Transformation

Sample Transformation

Last modified on November 6, 2026