On Jun 26, 6:30 pm, Allen <[EMAIL PROTECTED]> wrote: > Gandalf wrote: > > Hi every one I'm looking for a good alternative db to replace sqlite > > > I'm using pySQlite3, And I tried to translate very big database from > > Mysql to sqlite. > > I generated through PHP a python script that insert 200,000 records > > to my sqlite db and took me more then 5 hours and managed to insert > > only 100,000 records. > > I have almost million records so I need a better solution. > > > thank you > > I don't know if this will help, but using transactions dramatically > speed things up. I've only played around with inserting records from > large CVS files in C++, but outside of a transaction it creates an > automatic transaction every time an update is made and takes forever, > but if all the updates are inserted into one transaction and committed > as one, it is substantially faster.
In addition, if this is a one time conversion, look into using the bulk import functions. (.import FILE TABLE) Since you are coming from another database you know your constraints are satisfied, I would wait until all the data is loaded before building your indexes. HTH ... Jay Graves -- http://mail.python.org/mailman/listinfo/python-list
