Re: 1=1

Karsten Farell wrote:

> Joel Garry wrote:
> > I was hoping that wouldn't be the answer. But... it's the _last_
> > one...
> > [and I've seen others ending with AND 1=1) AND 1=1)]
> >
> > I just have some vague memory from years ago that it had some sort
> > of useful effect in rule-based.
> >
> > jg
> Maybe the 1=1 on the end is tacked on as a result of not selecting
> anything from a drop-down list on the UI (as the value of the default
> item). I suspect the cost-based optimizer throws away the 1=1 clauses
> anyhow (no matter where they are).
>
> You might be remembering the old tricks we used on the rule-based
> optimizer (did it really optimize anything?) to avoid index usage - eg,
> select column+0 or column||''. Perhaps they still work with CBO (unless
> it has grown so smart it knows when to ignore us).

There are times when using 1=1 makes a significant difference in the execution plan selected by the CBO. And thus, it can dramatically improve performance of a query.

As to why? I can't say. But I had a query once where dragging DUAL into it provided the same improvement.

Go figure. Not doubt Howard, Jonathan, Tom, and other can. I can't.

Daniel Morgan Received on Thu Oct 10 2002 - 13:28:00 CDT