Hi
I had a similar issue a few months ago, only it was several hundred thousand 
records rather than millions. I found that dividing the job into  smaller 
transactions did definitely speed things up considerably. In my case I think I 
limited each transaction to 20,000 records.
I would definitely agree that you don't want to spend too much time 
experimenting with lots of different methods; diminishing returns and all that. 
Do test runs of, say, 50,000 records and then just go with the first that seems 
to provide "reasonable" performance.
Regards
Nige


                                          
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Reply via email to