When using 'Net change' to update all orders which have been updated since the last run, incidentally far to many records are processed which have not been updated. Does anyone know why this happens?

In a map we update the salesorders with data from a csv-file. In the next map within the same solution 'Net change' is used to select all orders which have been updated since the last run. For these orders, all orderdetails are 'for-eached' checking a custom status. We are using 'variables' to store the lowest custom status. The order status is updated according to this latest status.


The solution runs every 2 minutes.

If the issue occurs, the Execution History looks like this:


6-11-2018 1:13 PM    00:01:28  8.5  Completed Successfully     746  0  0

6-11-2018 1:04 PM    00:06:27  7.3  Completed Successfully   2830  0  0

6-11-2018 10:24 AM  02:38:09  7.7  Completed Successfully 73133  0  0

6-11-2018 10:18 AM  00:03:49  3.2  Completed Successfully     739  0  0 


As shown, the run at 10.24 AM was 100 times the run at 10:18 AM while there were only a few extra orders imported.


It seems that 'Net change' did not use the correct date/time and processed too many records....