Re: [PyDB2-discuss] DB2 performance

[<yu...@ad...>]

Quoting Preston Landers <pla...@jo...>:

> On Fri, 27 Feb 2004, ??? wrote:
> 
> > Hi~ everyone
> >
> > As you can see in the source code, PyDB2 is not suited for performance yet.
> > The purpose of the first code I wrote was 'let's make it do something in
> hurry'.
> > There's a lot of things to improve to speed up the operation and should be
> coded in C.
> >
> > Welcome!
> >
> > I want to see the PyDB2 be speed-demon :)
> 
> 
> Hello everyone,
> 
> I'm curious to hear everyone's ideas on speeding up this driver.  I've
> only spent a small amount of time looking at the source code so far, but I
> don't have any ideas yet on speeding it up.  Then again I am not
> intimately familiar with DB2's low level protocol.
> 
> By the way, I have not yet checked in my enhacement to support the
> SQL_SUCCESS_WITH_INFO status code... I believe you had suggested there was
> a better way to handle storing the warning message, but I never quite
> understand what you wanted me to do with it.
> 
> thanks,
> Preston

I'm not really sure about the performance issue - I mean the DB2 api is fairly 
simple - prepare, execute and fetch (which is exactly what is being done in 
the module at the moment).  There are issues that I encountered, however.  One 
issue, for example, is that the connection and cursor objects are defined 
inside the c module itself which causes bad ref. counting management.  
Since the issue of a new implementation now arises I think that we could take 
a look in the MySQLdb code and learn how things are done there (module 
objects' definitions, error handling, various python optimizations, etc.)
What do you say?

-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/