Why are version 6 views so sluggish? How to impr

From: Barry Roomberg <barry.roomberg_at_compudata.com>
Date: Sun, 12 Jun 94 16:18:00 -0500
Message-ID: <12a.2619.905_at_compudata.com>


-> *** Quoting Keith W. Johnson to All dated 06-07-94 ***
-> > I support an old version 6 database on a small VAX/VMS machine. I
-> just > removed some common logic from several SQL*plus batch jobs and
-> created > a view that performed the same logic. The original SQL was
-> then modified to > include the view in a join. So... same logic, just
-> implemented differently.

-> Quite -- one wonders -why- you did such a thing, but we'll leave that
-> question aside for now...

Sounds like to remove dupe logic from batches, usually a GOOD idea.

-> > Are there ways to tweak the version 6 database so the view was not
-> such > a pig?
->
-> DO NOT USE A VIEW. It is not at all appropriate, given the
-> circumstances.

Not necesarily true. I recall that the orgianl V6 was horrible when attempting this type of logic on views, but later versions pushed the "where" criteria into the underlying view selects, when gave some incredible performance gains. Received on Sun Jun 12 1994 - 23:18:00 CEST

Original text of this message