On Tue, Jun 17, 2008 at 12:24 AM, Richard Broersma <[EMAIL PROTECTED]> wrote: > Anyway, here is a link discussing a generalized vertical partitioned > view. Perhaps it can give you some idea to get yourself rolling. > http://archives.postgresql.org/pgsql-general/2006-12/msg01119.php
Thank you very much for pointing this out! I am somewhat disturbed by the example in that thread of a "partially executed update" resulting from the obvious way to write an update rule for a view. I guess I need to be calling a function to do it instead, but this again seems somewhat kludgy. Perhaps rules are not as wonderful as they seemed when I first encountered them. > (I wished that the postgresql update rules where executed as > serializable transactions, that way if one of the joined tables in > the view was updated during your change, it would though an > exception rollback your update instead of writing over the other > persons work.) This is also disturbing! I'm not completely clear on what sort of overwriting can happen, though; could you give an example or a link? Are there any plans to fix these problems? In any case, it seems as though these sorts of caveats should appear in the documentation. Mike -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
