As others have suggested, transactions, create a single file and just copy,
or throw it in memory (And as an extra thought to creating it in memory is
once created, throw it to disk via the Backup API then --if needed-- work
off that).
Another option, but probably not the best idea, is each time a table is
going to be accessed, use a "CREATE TABLE {tablename} IF NOT EXISTS....".
This way the table exists when the call is made to do something with that
table. You'd probably have to comb through your code to find out when the
tables are to be accessed. Perhaps a function that calls the above SQL
statement, keep a list/collection variable hanging around that will keep
tabs of when a table was made during that session so you're not trying to
create the table EVERY call. The perk is that you might not have to have
all 1000 tables so the DB would probably be a tiny bit smaller? Ok, not a
BIG perk if you're dealing with 1000 tables.
My suggestion is a* last ditch option*. It is a lot of work to go through
code and re-test what already should be known to work with this new method,
but I figured I'd throw out another option.
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
