Subject Re: Basic tests shows very poor performance
Author Svein Erling Tysvær
Hi Toon!

I was surprised too when I read your original post. I'd recommend you
to try the same test with Firebird using random rather than duplicate
data. Inserting 150000 identical records in a table isn't a sensible
test, 300000 different records (possibly with a few duplicate ones)
would be more realistic. And it is generally recommended adding the PK
at the end of indexes with Firebird to improve the selectivity if
there are many duplicates in the indexed field.


--- In, grostoon wrote:
> Of course, I know that my test is not representative of real life
> access to a production database; this is not what I had in mind.
> No, I was just surprised that this basic scenario took so much time. 3
> or 4 minutes of elasped time, well ok; but nearly one hour; yes it
> surprises me.