On Wed, Sep 24, 2014 at 10:52 AM, Rob Golsteijn <rob.golste...@mapscape.eu> wrote:
> > Sqlite decides to create an AUTOMATIC INDEX (time complexity O(n log n)) > which it then uses to iterate table aaa1. This index is not re-used for > anything else (it can't be re-used since 'type' is not used anywhere else) > so only the traversal of table aaa1 benefits from this index. However, I > think, a full table scan of aaa1 (time complexity O(n)) would always be > faster, since for creating the index it has to read that entire table > anyway. > > I'm using sqlite version 3.8.4.3 > Thanks for the report. That problem appears to have been fixed in SQLite 3.8.6. -- D. Richard Hipp d...@sqlite.org _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users