Then we perform the changes for those rows. We use the Map Operation to filter on a specific OPCODE and at the same time change the OPCODE to normal so a query can be added downstream.

But for both a Map Operation Trnasform can be used. The problem is, neither is there a function to identify the OPCODE nur can you connect the output of the transform to a query. The only direct solution, yet an ugly one, is described here.Ī Table Comparison Transform is outputting the row if something changed and each flag will have an OPCODE of insert or update (or delete if configured). Either way, all have the character of a workaround. Or you use a Slow Changing Dimension Type 2 (SDC2) instead where the first recod holds the create date and you can see all change dates in the subsequent rows. you could ignore those two fields in DI and have a database trigger instead. This is something where there is no good solution for in Data Integrator. Often it is requested to have separate columns of when a row got inserted the firt time - the create time - and when it was updated.

How to use confluence to create a date