Hi Horst,

one possibility is that somewhere in your ASN file there is a redirection for *.SQL files
to a location which does not exist or the user has not enough privileges..

So uniface is not able to write the output file.

Hi Uli

Thanks for the advice, but as I have seen this cannot be the cause of the problem. Actually I am able to generate a script, so it is not a redirection issue. Without any Ref-Integrity "defined" the script is empty as expected, but the error message at the end (of the creation process in the message line) still shows up. Doing the same thing in a different MODEL creates the empty file without this error message. The original cause of my problem is, that during the generation of the script, only a part of the RI-Constraints are created and the rest is not created (any more, it worked before). Now I am looking into all possible causes which could lead to this strange behaviour. The Ref-Integrity itself seems to be OK, I can create them all in Oracle manually. But obviously we have introduced an error into our datamodel with some recent enhancements which causes the create utility to stop at a certain point during the creation of the ref-integrity constraints. Without knowing the error details and/or how the constraint creation script works in general, it is pretty difficult to resolve this issue.
So actually I am looking for
a) has someone had this problem before,
b) are there any logs for the constraint creation utility I could check and
c) is there a way to search for relationships in Uniface which could cause such a problem. (for example in the database itself)

Thanks anyway for your help, I keep trying to enable the constraints one by one until I find from when on it is "screwing" up.

Horst

Hello Horst,

just some brainstorming contributions

does the sequence of SQL statements in the script gives you any hint whuich entity to look into?

When you analyze the model in question, do you get any obscure messages which may help?

Do you have a relationship to another model?

Have you changed a datatype of a FK field which does not match the PK field any more ??

Just for experiments:

when you use case unload to dump the model in question with all the relationships;

rename the model in the CIF file

and CASE load it again: will this make any differences in generating the SQL ?