Re: performance of model instgance save()

2011-05-17 Thread Lucian Nicolescu
I had at similar problem at some point and found that the fasest way was to generate raw sql insert/copy commands - if you're still experimenting you might want to try that as well. On Tue, May 17, 2011 at 12:39 AM, Brian wrote: >> > I'd think you'll find it's significantly more than a factor of

Re: performance of model instgance save()

2011-05-16 Thread Brian
> > I'd think you'll find it's significantly more than a factor of 2. For bulk > > inserts, raw SQL is often several orders of magnitude faster. > > You might want to check outhttp://pypi.python.org/pypi/dse/. My tests > using postgres showed a 3Xperformancegain on inserts compared to > using the o

Re: performance of model instgance save()

2011-05-11 Thread Thomas Weholt
On Wed, May 11, 2011 at 10:21 AM, Daniel Roseman wrote: > On Wednesday, May 11, 2011 7:50:16 AM UTC+1, Brian wrote: >> >> Sure - it's just that bit more convenient and readable to use the ORM >> approach. If I get the chance, I must try substituting direct calls to >> psycopg and see if that makes

Re: performance of model instgance save()

2011-05-11 Thread Daniel Roseman
On Wednesday, May 11, 2011 7:50:16 AM UTC+1, Brian wrote: > > Sure - it's just that bit more convenient and readable to use the ORM > approach. If I get the chance, I must try substituting direct calls to > psycopg and see if that makes it significantly faster again. I'd guess > there's maybe an

Re: performance of model instgance save()

2011-05-10 Thread Brian
Sure - it's just that bit more convenient and readable to use the ORM approach. If I get the chance, I must try substituting direct calls to psycopg and see if that makes it significantly faster again. I'd guess there's maybe another factor of 2 to be had. -- You received this message because

Re: performance of model instgance save()

2011-05-10 Thread Venkatraman S
On Wed, May 11, 2011 at 4:51 AM, Brian wrote: > Having hauled myself a few feet up out of the abyss of ignorance, I can > answer my own question (which might be of benefit to others getting started > with django). To clarify the problem, I have a standalone script that > imports the django settin

Re: performance of model instgance save()

2011-05-10 Thread Brian
Having hauled myself a few feet up out of the abyss of ignorance, I can answer my own question (which might be of benefit to others getting started with django). To clarify the problem, I have a standalone script that imports the django settings and uses its ORM pleasantness to populate one of