2008/12/1 Igor Tandetnik <[EMAIL PROTECTED]>: > Try this: > > select n from map > group by n > having > count(case when m=3 then 1 else null end) != 0 and > count(case when m=5 then 1 else null end) != 0 and > count(case when m=7 then 1 else null end) = 0; > > Having an index on map(n) should speed it up.
Thank you very much, Igor. I would have not thought of that. This is a nicely predictable single linear scan. Still not awfully fast, but it will have to do. Using an index on map(n,m) seems faster. This may be because the m values are in the index and there is no need to access the table. Cheers, Jos _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
