Re: Unknown SQL

In article <u7k82yvskr.fsf_at_sol6.ebi.ac.uk>, Philip Lijnzaad says...
>
>On Thu, 31 May 2001 03:12:21 +0200,
>"Carl" == Carl Rosenberger <carl_at_db4o.com> wrote:
>
>>> of your experiments, I do not know what you did wrong to cause the
>>> performance problems you saw. My guess is you did not have the DBMS
>>> calculate the statistics properly.
>
>Carl> Vadim has just posted the solution. The optimizer was not able to
>Carl> understand GROUP BY statements.
>
>If I read his post correctly, this is not a problem with the view per se, but
>with the fully expanded query being too complex. I.e., the equivalent (fully
>expanded) query would have exactly the same performance. Is that correct, Vadim?
>

Correct, blame optimizer, not views:-)

And even that was not the whole story. They added some "complex_view_merging" parameter working with Cost Based Optimiser that relaxes the forementioned restrictions on view merging. It merges some complex, aggregate and set operator views. Not surprisingly, it adds some cost based heuristics. There is also "_push_join_predicate" parameter... Received on Thu May 31 2001 - 18:23:09 CEST