If you have a routine that uses a lot of data or returns a lot of data, best practice is to pass clusters to the routine.

Clusters are passed by reference, which means anything about a cluster passed into a routine can be changed by the routine. This includes adding new rows to a cluster array, or changing the values of cluster fields.

Below in the editor is an example of a routine that calculates tax. We will pass into the routine a cluster that contains the amount and the tax rate. The routine will return the tax due.

A routine having a cluster passed to it has full access to the cluster for both reading and writing. This includes being able to add new rows to the cluster, performing collect/end collect operations, etc.

Simplified Cluster Parameters
Just like simple parameters, cluster parameters can also be simplified if the root cluster contains data. In this example, the root cluster is called taxes and is both a parameter name and a cluster name.