Re: performance is slow

I have some results. I create the cursor and keep it active until I have
inserted all of the information that I have and done all of my updates
at which point I do a commit and then close the cursor. It has improved
the 44 name-value pair entries by an order of magnitude but only halved
the blob insertion. All in all it is a 5x improvement for the whole
process. It is at least becoming tolerable.

Any more suggestions on how to speed things up? The one thing that I do
not like is that I see the network usage sky rocket during this time. Is
there a way that I can have a pipe connection instead of network?

On Wed, 2009-06-10 at 11:17 -0700, Tim Roberts wrote:
> Al Niessner wrote:
> > The main thing that I think I am doing wrong that I want to change is
> > that I am doing a commit after after insert. I am looking at changing
> > that to a bigger more "transactional" item. Not sure if this will
> > improve the overall time, but it seems like a more DB way of doing
> > things.
> >   
> 
> That is the key right there.  If you change your code so that you do not
> commit until your whole series of inserts and updates is complete,
> you'll find that the performance is quite acceptable.
> 
-- 
Al Niessner

I have never found the companion that was so companionable as solitude.
- From Walden by Henry David Thoreau

The universe is indifferent, and life is brutal; however, it is man's
choice of behavior that makes them malevolent rather than benevolent.

Some will fall in love with life and drink it from a fountain
That is pouring like an avalanche coming down the mountain.
- From the song Pepper by the Butthole Surfers