I have a table with two variables, say A and B (both integers). The table is rather large - around 2.9 GB on disk. Every combination of (A,B) occurs only once. I am creating a unique index as CREATE UNIQUE INDEX ABidx ON abtable (A,B) It seems that the (A,B) index is created much slower than the (B,A) index. I am wondering about the reason for this. My - very limited - understanding was that sqlite needs to search over the whole database for every (A,B) combination. Could someone give a quick pointer to where the index strategy is described? Would this be quicker if I fit the whole database into memory?
Kasper _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users