Data Processing Pattern
The pattern in which IMan processes data becomes relevant for transforms involving hierarchical data.
Example
In the Map and Filter transforms, IMan starts with the innermost child and works outwards to its parent before moving on to the innermost child of the next node.
Dependency
Since the processing pattern works this way, it is possible to create parents that are dependent on the results of their children, but not possible to create children that are dependent on their parent's result. The only way to create a dependent child transaction, would be if two or more transforms are used.
example: Parent Dependency in Filter Transform
A filter is placed on child records which, when evaluated, causes each of the child records to be removed from the dataset. A filter is then put on the header records to delete the header record if it has no child records.
Result: The header records are now reliant on the result of the child record evaluation.