Hi all,
We create our dataenvironments programmatically in our VFP6.0 -application. Our dataenvironment consists of separate cursors, which are nowadays based on pre-defined (mostly) and runtime-defined views. Technically everything works fine, but our problem is a serious lack in effectiveness when using a view instead of a table as a cursorsource of the cursor. It namely seems, that if have a wide table (about 90 columns) and we save our data to it by tableupdate-function, the saving time is about 7 times longer when there is a view between the cursor and the table. Does somebody have any ideas what is going on and how to get the process faster?
I have already tried following things without any bigger success:
- change the bufferingmode of the cursor (from our normal 5 (optimistic table buffering) to every possible value)
- change 'SQL Where Clause includes'- and 'Update using' -definitions in the View-designer of VFP
- change the 'nrows'-option of TABLEUPDATE-function to all possible values
Reducing the amount of returned columns and rows improves the situation, but that solution is not practicable in this special case.
Best regards,
Ilkka Koskinen
Email:
ilkka.koskinen@tieto.com