I have a program written using SF that loops through tens of thousands of records and must check multiple tables and perform calculation, string builds, and other perform other parsing for each record. I am using "raw" datasets, datatables, and dataviews due to the fact that I am accessing DB2 files. The results of the raw processing are then stored to a BO within SF.



This process is very very slow (about 2 minutes) during which the application is completely locked up, if it is minimized it stays minimized, nothing can be interacted with, etc etc. The bad part is this process occurs every 3 minutes. These are the requirements of the app.



Would introducing threading help me, or am I just screwed given the scenario presented?



Any insight appreciated.



Also, did you guys ever discuss internally fixing the fact that DB2 tables are not properly enumerated in the BO Mapper using OLE?



Thanks.

Keith Chisarik

Absolutely!  This is the exact type of thing that should be run and processed on a thread.

Not yet.  It is on our "To-do" list though.  We have had some server issues recently and need to get DB2 loaded for testing.

Freaking awesome!!



I have three threads now, one dedicated to the UI, and one for each of the two major processes.



You would never even know they were running now except for my nifty little progress monitors



Again, thanks for the nudge in the right direction.



Keith

Keith Chisarik