# Coalesce data

> Supported in: Batch, Faster

Operation to reduce the number of partitions. If you have 1000 partitions and you coalesce to 100 there will not be a shuffle, instead each of the 100 new partitions will claim 10 of the current partitions. If a larger number of partitions is requested, it will stay at the current number of partitions.

**Transform categories**: Other

## Declared arguments

* **Dataset:** Dataset to perform coalesce on.<br>*Table*
* **Number of partitions:** Number of partitions to coalesce to.<br>*Literal\<Integer>*