Search results

DEFINE VARIABLE cChanges AS CHARACTER NO-UNDO. DEFINE TEMP-TABLE ttRechnung LIKE Rechnung. FIND FIRST Rechnung NO-LOCK. CREATE ttRechnung. BUFFER-COPY Rechnung EXCEPT Rechnung.RG-DAT TO ttRechnung. BUFFER-COMPARE Rechnung TO ttRechnung SAVE RESULT IN cChanges. MESSAGE cChanges VIEW-AS...

If you define "changed" as a character you should get a list of the changes. I think.

Just to make sure I understand - you want to ensure at the point of locking the record in the new buffer that it hasn't been changed by another process? If so, then you can use the BUFFER-COMPARE keyword to compare the find-customer to the upd-customer and at that point make a decision as to...

One thread will suffice. I've closed this so there's no confusion.

The obvious question is why you want to do that. What are you trying to achieve?

It's superbly good for little snippets of code. I'm with Stefan - if it has meaning it goes into SCM. I use it in 12.2.6 without any issues like you mention. It's still dysfunctional as Stefan says but I never lose code. I don't remember it being a problem in 11.6 either but that was a very...

It would help if you would give us the errors you receive rather than just saying it doesn't work. In this case I guess it doesn't work because you need to define the output location: output to value (vCSVFileName) [append]. And you'll need to close the output as well once done.

Another one to watch out for is changes to the length of an extent.

I've worked under numerous folks over the years, all with different approaches, but the one that caused the most food for thought was the policy that all schema change requests had to also come with sample data in a .d file. This had a number of purposes. 1) The developer has actually tested...

Aren't they in the df of the database? Can't you just therefore create an incremental df to load?

Hi @Ross55, this forum has nothing to do with Pervasive databases. I doubt anyone will be able to help.

It does indeed work. Although I agree with @Stefan that the complication is bad.. FUNCTION MyFunc RETURNS INTEGER (): MESSAGE 111 VIEW-AS ALERT-BOX. RETURN 10. END FUNCTION. DEFINE VARIABLE i AS INTEGER NO-UNDO. DO i = myfunc () TO 1 BY -1: END.

Your appserver will have a broker and server log file. These are a good place to look for errors.

Please can you post a more complete code sample as a starting point.

In reality assigning the NUM-ENTRIES () to a variable probably makes very little difference, but Tom is correct, that one probably should. I'll leave the religious wars on keyword casing to others. As well as the variable naming conventions. I write my code based upon the conventions of the...

o_O

I was a bit busy earlier, but I have had time to write a more efficient solution for the Sports DB as an example: DEFINE VARIABLE cCountries AS CHARACTER NO-UNDO INITIAL "USA,Finland,France". DEFINE VARIABLE i AS INTEGER NO-UNDO. DEFINE VARIABLE cCurrentCountry AS...

I agree the first will be better, and assigning the ENTRY() result to a variable will make a difference.

Both options will almost certainly result in a whole index read of the Order table. Functions in query predicates are typically hard to resolve and so you end up with poor performance. You could use an OPEN QUERY or a Dynamic Query to solve the problem more efficiently.