From your description I take for granted, that you know WHICH child records have to be cretaed for WHICH master record, so I guess there are basiacally two possibilities.
- All data is known at runtime:
- Create a dataflow, which populates the master table
- Cretae a second dataflow, which populates the child table and connect it to the first dataflow via Succeed constraint
But basically this sounds bit too easy - guess you tried this out already? So here comes possibility 2: the parent table creates IDs, which have to be referenced by the child records:
- Again we start with one dataflow in order to populate the master table
- then we add a second dataflow - again attached to the first one via succeed constraint
- in this dataflow we add a lookup, which checks the master table for the corresponding IDs
- This ID is then written to the corresponding column in the child table.