Re: Bulk data insert

Agreed, not sure my suggestion will help you. You can have a look at turning off transaction, that might make your code faster but I'm really not sure. ( Check-out those from django.db import transaction @transaction.commit_manually ) - Benjamin On Thu, Aug 20, 2009 at 9:58 PM, Doug

Re: Bulk data insert

On Thu, Aug 20, 2009 at 10:40 PM, Benjamin Sergeant wrote: > > (the answer was already sent, raw SQL, anyway) > > (with postgresql) I would try to load just a small bit of datas with > the slow create django object / save it process, then do a pg sql > dump, and look at the

Re: Bulk data insert

(small update, if you try what I just suggested and use postgres (lots of if) you want to generate SQL and load with psql and not pg_restore) - Benjamin On Thu, Aug 20, 2009 at 7:40 PM, Benjamin Sergeant wrote: > (the answer was already sent, raw SQL, anyway) > > (with

Re: Bulk data insert

(the answer was already sent, raw SQL, anyway) (with postgresql) I would try to load just a small bit of datas with the slow create django object / save it process, then do a pg sql dump, and look at the sql that pg dump generated. Then code a python script that generate that. And then do a pg

Re: Bulk data insert

On Aug 20, 2:50 pm, Alex Gaynor wrote: > On Thu, Aug 20, 2009 at 1:46 PM, Abiel wrote: > > > Is there an efficient way to use Django models to load a large number > > of records into a database without falling back on raw SQL? Creating a > > large number

Re: Bulk data insert

Thanks for the feedback, I will go ahead and use raw SQL. Hopefully bulk inserts will be a feature in a future version of Django, especially given the rising importance of sites oriented towards large volumes of raw data. Abiel On Aug 20, 2:50 pm, Alex Gaynor wrote: > On

Re: Bulk data insert

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Abiel declared: > Is there an efficient way to use Django models to load a large number > of records into a database without falling back on raw SQL? Creating a > large number of model objects and then saving each one individually is > very slow (I

Re: Bulk data insert

On Thu, Aug 20, 2009 at 1:46 PM, Abiel wrote: > > Is there an efficient way to use Django models to load a large number > of records into a database without falling back on raw SQL? Creating a > large number of model objects and then saving each one individually is > very slow

Bulk data insert

Is there an efficient way to use Django models to load a large number of records into a database without falling back on raw SQL? Creating a large number of model objects and then saving each one individually is very slow (I imagine Django is running INSERT and COMMIT each time?). Thanks very