[dev] datatree performance (again)

[Eric Jon Rostetter]

Quoting Jan Schneider <jan at horde.org>:

>> Output of explain's extra column shows that MySQL has to use operations
>> that are known to be expensive in terms of performance: Using where, Using
>> temporary, Using filesort
>
> That's wrong. "using where" and "using filesort" are the fastest
> expectable operations when doing queries with WHERE and ORDER
> statements. "Using temporary" is also expected because we use
> different column orders in the ORDER BY and GROUP BY lists. Perhaps we
> could achieve the same results without requiring temporary by
> reordering the columns though.

I'm running a FRAMEWORK_3 checkout with postgresql and not seeing a
terrible hit (though the postgresql load is higher than before; my load
average went from 0.0 before to about 0.25 now).  Since this is on a dual
CPU P4 XEON 2.4 machine with hyperthreading enabled, I have some spare cpu
with such a low load...

Anyway, could it be this is a MySQL problem, which doesn't affect (at
least some) other databases?  Maybe take a poll as to what databases
people seeing this issue are using, and which databases people not seeing
the issue are using?

> Jan.

-- 
Eric Rostetter
The Department of Physics
The University of Texas at Austin

Go Longhorns!