On 2/8/19, David Raymond <david.raym...@tomtom.com> wrote: > Non-scientific "let's just try it" results > > Short version: > Original file had been vacuumed already as the last thing that had happened > to it. > File size: 20,467,359,744 > > sqlite> vacuum into 'vac_into.sqlite'; > Run Time: real 589.577 user 222.941029 sys 57.829571 > > sqlite> vacuum; > Run Time: real 1429.063 user 236.325915 sys 199.322478 > > Synchronous was off, journal mode was normal rollback journal.
Fair enough. I wasn't thinking about the overhead of journaling when a normal VACUUM copies the new database back over top of itself. It is having to do two complete copies of the database, not just one. So you may well get more than 2x better performance. -- D. Richard Hipp d...@sqlite.org _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
