>>>You did say at one point the whole thing took much longer and then suddenly it was much improved. I didn't notice if it was exclusive or not. Network traffic could be the culprit.
>>
>>Something happened with the PC where this process was running, I think. I have no explanation for 2h. of running this process.
>>
>>>Now, a difference of only 5 seconds seems to disprove the chunking theory. It would fit my thinking of there being only a difference in the number of executing lines of code. Small difference, but there. How many records are we talking about?
>>
>>Just 2046 records to process.
>
>Then there's no need to spend time optimizing this.

I think I see what's going on. I'm studying right now Batch Services code and see code similar to this in various places:

DO WHILE lcErr == "Error Found!"
			TRY
				lnAttempt = lnAttempt + 1
				IF lnAttempt != 100000
					OPEN DATABASE (thisform.cDatabase) SHARED
					lcErr = "Error Not Found!"				
				ELSE
					lcErr = "Error Not Found!"									
				ENDIF	
			CATCH TO oException1				
				lcErr = "Error Found!"				
			ENDTRY
		ENDDO