Cliff, happy to help. All dataflows in Data Services attach an operation code to every row of data. By default, this is the NORMAL operation code (which acts like an insert). The Table Comparison transform:
1) Joins incoming and lookup data by the input primary keys
2) Compares rows in the comparison columns
3) If no change is present, the input record is not sent out of the table comparison transform at all
4) If a change is detected in the comparison columns the operation code is changed to update
5) Assuming you connect the Table Comparison transform directly to the target table, the keys of the target table are used to perform the update, which are done 1 by 1.
In short, there is no need to create another step to perform the updates.
Let's say we have this data:
Source Target
1, ABC 1,ABC
2, DEF 2,EFG
The result of running this data through the Table Compare would be 1 row with an operation code of UPDATE. The first record, since it is identical, would be removed from the output of the Table Comparison transform.
You are correct, no History Preserve transform, no Slowly Changing Dimension Type 2 ![]()
Alexei







