A transform definition specifies the sequence of operations applied to the data during
a batch data transform. Definitions appear as nodes in the data flow canvas. You can combine
multiple definitions to produce a single, unified output.
Required Editions
| User Permission Sets
Needed |
| To create batch data transforms |
Permission set:
|
A transform must have at least one definition, but it can support up to five. When writing to
an output object, Upsert (or "merge") Mode processes only new and updated records. This
feature optimizes resource usage and limits unnecessary reprocessing. Upsert Mode is typically
used with multi-definition transforms.
Consider these limitations when using multi-definition transforms:
- If you have multiple transform definitions for a single transform, only one definition can
be scheduled to run at a time, but a flow can trigger any of the definitions to run.
- Definitions within a single transform can write to the same output object. Different
transforms can’t write to the same output object. For example, when multiple definitions
belong to the same transform, a single definition can write to the same output object.