russt at releasetools.org
Fri Mar 1 09:57:12 PST 2013
I agree with you that JDBC.pm is not a practical idea for production database work, though it works remarkably well.
This is more of an academic exercise, which was the reason Tim Bunce originally wrote JDBC.pm - to help explore the java model as a means to inform future DBI development.
Think of this as a pure exercise in optimization, given a constrained problem.
At 9:35 AM -0800 3/1/13, Francisco Obispo wrote:
>On Mar 1, 2013, at 9:22 AM, Russ Tremain <russt at releasetools.org> wrote:
>> Geez, lots of opportunities for optimization here, where should I start?
>I would suggest, by removing all Java-related code.
>> What would you do to improve the performance here?
>Learn and improve your code using DBIx::Class
>> I will post what I actually did in a couple of days to give you a chance to think about it.
>> Hint: anything derived from $rset is handled by Inline::Java, which means at least one write/read cycle to a socket connected java VM.
>I have seen programs written in pure Java, fail miserably on high performance database applications, mostly IO bound on context changes. If you really like the java code, I would suggest you use java, and not try to mix it with perl.
>Director of Applications and Services - ISC
>email: fobispo at isc.org
>Phone: +1 650 423 1374 || INOC-DBA *3557* NOC
>PGP KeyID = B38DB1BE
More information about the SanFrancisco-pm