On 18 Feb 2002, Jeff Zucker wrote:

> Matt Sergeant wrote:
> >
> > Some benchmarks:
> >
> > First DBD::CSV latest from CPAN:
> >
> > Using /root/.cpan/build/DBD-CSV-0.2001/blib
>
> The relevant thing for the speed of DBD-CSV is the version of
> SQL::Statement you are using.  If you have the newer perl version
> (version 1.002) , I have only recently begun working on optimization. A
> more reliable speed test will be with using the older XS version
> (version 0.1021).

This was 0.1021 of SQL::Statement.

> I would also be very interested to hear other comparisons than speed.

Yeah, I'd like to get it all working first ;-) One area I expect it to
slow down on is large result sets, simply because of the aforementioned
problem with how SQLite deals with results. However it's not tragical, as
it's just a char** I have to loop over (though it frustratingly isn't null
terminated!).

> It sounds great.  I'll try it as soon as I can.  Congrats!

Thanks! It was actually easier than I thought it would be, though of
course SQLite did the hard work for me.

-- 
<!-- Matt -->
<:->Get a smart net</:->

Reply via email to