In the call to create index, it looks like you're creating a duplicate index on [id], which I believe references the default btree id (also called rowid.) If you change that line to:
CREATE INDEX [by_id] ON [profile_data] ([profile_id]); your index should be used (though whether indexes get used with select distinct clauses, I'm not sure.) Ian On Nov 8, 2007 4:30 AM, Maxim V. Shiyanovsky <[EMAIL PROTECTED]> wrote: > Suppose I have simple table: > > CREATE TABLE [profile_data] ( > > [profile_id] INTEGER, > > [version] INTEGER); > > CREATE INDEX [by_id] ON [profile_data] ([id]); > > > > Why > > sqlite> explain query plan > > ...> select distinct(profile_id) from profile_data; > > 0|0|TABLE profile_data > > > > Does not use index? > > ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------
