Re: Oracle vs MySql Performance

Johne_uk wrote:
> Hi,
>
> I'm looking for a few opinions on an issue we are experiencing.
>
> My company currently uses a 3rd party application. Up until a few
> weeks ago this application consisted of Java code and a database layer
> that used MySql.
>
> Performance on MySql was adequate but as our primary DB platform is
> Oracle we influenced the 3rd part to introduce an Oracle version (NB:
> most of the sql code is stored in a java db layer - schema just
> consists of tables, views and indexes).
>
> To facilitate this process the vendor has used almost pure ANSI Sql so
> that the same sql can be used for both oracle / mysql with minimum
> conversion required.
>
> When the system was released we were surprised to find that the sql
> (and hence application) ran slower on oracle than MySql (the Oracle db
> server is also considerably more powerful).
>
> At a high level does this surprise anybody on this forum. Does Oracle
> struggle to execute ANSI sql compared to MySql.
>
> The orginal code was also written for MySql and converted so that
> might explain why is is running slower.
>
> Any thoughts would be appreciated.
>
> Tanks in advance
> John

When you wrote: "the vendor has used almost pure ANSI Sql" you answered your own question. Not that there is anything wrong with ANSI SQL per se but rather because the truth is your vendor knows almost nothing about Oracle and used mediocre knowledge to build something decidedly mediocre.

If something other than a heap table would have helped they didn't use it.

If something other than a B*Tree index would have helped they didn't use   it.

If array processing would have helped they didn't use it.

Tom Kyte has repeatedly warned against this. Of course they likely don't even know who Tom is.

-- 
Daniel A. Morgan
University of Washington
damorgan_at_x.washington.edu (replace x with u to respond)
Puget Sound Oracle Users Group
www.psoug.org