Moshe Jacobson <moshe@xxxxxxxxxxxx> wrote: > I can provide any other information you want, just let me know. The best thing would be to show a *self-contained* test case -- that is, starting with an empty database create tables and (if needed) the view and show a query which demonstrates the problem, all in the form of a SQL script. Perhaps you only need to or three tables to demonstrate the effect; if you need more, that in itself is a valuable clue. If a developer can make the issue occur on their own machine, they can investigate much more effectively than without that. This is especially true when something works as expected in a number tests on a wide variety of environments; something must be different about your environment, tables, or queries -- but what? A self- contained test will make that clear. -- Kevin Grittner EDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
