Very nice ! the script list exactly what I need , thank you very much .
About the influence about dev team I agree too , but they keep saying that the corruption problem do not come from update or new script , another problem is that our new deployement system work with patch , before we did backup/restore for each major update , that was a little constraint but there was less errors , at least corruption problem were discover before going too big .
I hope to find someday the cause of corruption with no doubt about it source , for now i am not experienced enough to do this , I keep working on my repair tool and after that maybe I will find out why .

Many thank for your help and Alan too , have a nice day .
regards ,

--- In firebird-support@yahoogroups.com, Svein Erling Tysvær <svein.erling.tysvaer@...> wrote:
>
> >I think you are 100 % right , the problem is that i am not working in the developpement team but in customer
> >support , I suspect some software update to cause the problem but I have no access on that , the only thing I can
> >do is to try to repair the best way for my client and I can't only validate and mend , I have also to put back the
> >records from a backup with new ID or the problem will not be solve completely .
> >
> >We have about 500 servers too, the structure is from 2 up to 20 clients , little workgroup under windows XP and 7
> >
> >I don't think the problem concern the firebird side , I know it's very strenght.
>
> You also ought to try to influence your development team somehow. Normally, it would not be OK for you to have 100 corruptions (assuming an average of 2 per week) when Alan hasn't experienced any corruption at all. Hardware problems and faulty UDFs used to be ways to make Firebird fail.
>
> The following may or may not help you:
>
> select i.rdb$relation_name, iseg.rdb$field_name
> from rdb$relation_constraints rec
> join rdb$ref_constraints rfc on rec.rdb$constraint_name = rfc.rdb$constraint_name
> join rdb$relation_constraints rec2 on rfc.rdb$const_name_uq = rec2.rdb$constraint_name
> join rdb$index_segments iseg2 on rec2.rdb$index_name = iseg2.rdb$index_name
> join rdb$indices i2 on rec2.rdb$index_name = i2.rdb$index_name
> join rdb$indices i on rec.rdb$index_name = i.rdb$index_name
> join rdb$index_segments iseg on rec.rdb$index_name = iseg.rdb$index_name
> where i2.rdb$relation_name = :MyTableName
> and iseg2.rdb$field_name = :MyPKFieldName
>
> It is not tested, I know it does prepare and that it is possible to get some results, but I don't know if they are correct or complete (I'm assuming one field for each primary key).
>
> HTH,
> Set
>