On Thu, Mar 13, 2014 at 11:06 PM, Richard Hipp <d...@sqlite.org> wrote:
> > Once you do that, you'll see that the opcode sequence is only slightly > different between the two. They should both run at about the same speed. > I doubt you'll be able to measure the difference. > > Actually a comparatively long (10,000,000 elements) CTE for random integer generation shows difference 20 vs 38 seconds. I suppose pure min should use linear search while "order by" one uses temporal b-tree (exlain query also hints about this). Sure unless sqlite has some detection of "order by limit 1" pattern redirecting it to linear search. Max _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users