On Oct 27, 2008, at 12:38 PM, Julian Bui wrote: > Thanks for the reply dan. > > You probably don't "need" clustered indexing as such, but this would > be >> >> the kind of case where it provides some advantages. You can get the >> same >> effect in SQLite by including all the data columns in your index >> definition. >> >> > Unfortunately, because I will be making very frequent inserts (every > 10-100ms), an index spanning all columns would be much too expensive > and > would hurt insert times.
Have you been able to verify this experimentally? Predicting performance is a tricky exercise. Dan. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users