Hi,

I am completely don't understand your situation. Please provide an
example if you can.

Gabor


olaf.kluge@... írta:

> Hi,
>
> It is a stored procedure. It goes through every single position (data record) of an order-table and compacts them if possible. Then inserts the stored procedure it into a table and a second stored procedure takes each position (after the first stored procedure has created all positions) (data records) and examines the needed material from the high-bay-storage. But if the data record is missing in the one table, the second stored procedure can not works completely. I have found the problem in the first stored procedure.
>
> the uncompressed size of the database is 45 megabytes and the database is one month online. In this condition the first stored procedure often forgets the last position (data record). After I had compressed the database (18 megabytes), the stored procedure attends all positions (. There is no apparent reason why it omits the position, even not, if I debug the stored procedure.
>
> each data record in both tables has a unique integer identification value. it is also not a great value in a smaller writing, it is always the ongoing nummer. Often it is written into the database, about 10 updates every second. The java application with access to the database does auto-commit. (not written by me) The database is also not createc by me, but it is now my problem :o(
>
> For fundamental changes, it is too late, the production has begun at the customer side. Almost no table has an primary key, but that should not be the problem, it works fine the most time and in these two tables a continuous number identifies the records.
>
> I have no idea, why it works fine and if the database has grown, it works not always properly.
>
> Has someone an idea?
>
> Thanks.