[V4RB2] Holy crap, what a speed difference!

[Ruslan Zasukhin]

On 4/1/05 6:56 AM, "Kem Tekinay" <ktekinay at mactechnologies.com> wrote:

Hi Kem,

> I am converting my project from the SQL method, using a Vcursor, to the
> Table method, using bitsets, and the difference is remarkable! Searches that
> took over a second or two before are practically instant now. And sorts are
> incredibly fast too.
> 
> Where before I was finding that the REALdatabase was similar to, or faster
> than, Valentina, now I find there is no comparison, but only if I use the
> Table method.

Well, I think in 2.0 SQL parser takes many time.
Because it is bigger. And Because it is not hand-made.

We use ANTLR. Its developer promise to summer release ANTLR 3.0
They say it will be much faster of ANTLR 2.0.

Another point. In fact we have not spend yet even an hour to OPTIMIZE
Valentina 2.0.  We work for now on correctness.


> The only place where I still find Valentina slower is in retrieving data
> from records in a loop, but perhaps the people at Paradigma can help me
> optimize this once I'm ready to release my project.

Loop on table I assume.

Here also we can try to make optimizations yet.


> BTW, I see that Valentina created an index on-the-fly when I sorted on a
> non-indexed field. This is a very nice touch, but I wonder how I can turn
> that feature off if I want to?

In future versions it should be able sort on non indexed fields.
In Mantis exists such feature request.

Although this will be slower of course.

-- 
Best regards,

Ruslan Zasukhin
VP Engineering and New Technology
Paradigma Software, Inc

Valentina - Joining Worlds of Information
http://www.paradigmasoft.com

[I feel the need: the need for speed]