Re: Help! Very bad SQL bug in 7.2.3.0.0

Tom Fielden wrote:
>
> Recipe for bad results:
>
> Join several tables sequentially like so: a-b-c-d-e
> Put a simple on 'e'.
> Observe that the query runs fast.
> Now put a simple condition on 'a'
> observe that the query is 100 times slower! Sheesh!
>
> This is so incredibly bad that Microsoft SQL server looks like a
> viable alternative.
>
> Please don't let me get sucked into the 'Gates' of hell! What is
> the work-around?
>
> Also, this seems like an extremely common type of query to run. In
> fact, a developer down the hall has an almost identical problem with
> the same effect.
>
> If you have any ideas, I would greatly appreciate hearing from you.
>
> Thanks in advance,
>
> Tom Fielden
> Mentor Graphics Corp.
Whatever you are doing in B is almost surely causing you to do a full table
scan somewhere along the way. You need to learn to use Explain-Plan that is
documented in the tuning documents. You can find out what you have done wrong,
whether or not you need a hint or index.

Loyal Received on Sun Oct 13 1996 - 00:00:00 CEST