In such cases a dozen years ago the "solution" was to recreate reproducible code showing the behaviour with all unneccassary stuff left out and firing that off to Redmond as bug - after testing on other machines or giving it to coworkers for their take on it. Fixes from Redmond are more than improbable,but the process could give us something to run and not only think about...

BTW: those repeated replace commands on the same "in cAPP" cursor are cruel and unusual punishment for your CPU at least, in case of a materialized (as in disk-based) cursor for your disk as well ;-)


>You are absolutely right Tore. It doesn't make sense but it's the only suggestion that made the isnert work.
>To fix something one needs to know what needs to be fixed ;-)
>I never experienced a problem like that one before. I reallt don't know what I sould do to fix this the logical way.
>
>>Your explanation doesn't make sense to me, I suggest you spend some more time and fix whatever is wrong in your code.

>>>Problem solved... but I still don't get it.
>>>Solution was to put a go top in the cursor where I wanted to do an insert.