Data is stored in pages, and if you have a large blob as the first column, the other data is being pushed out to other pages (I think they are called "internal pages" in the docs). So to read the other data, sqlite has to first read the row's header page, then find & read the page that the other data is on.
Seems more efficient to have the small columns close to the key, so once slqite finds the row, the small columns are right there in that first page. That's not unique to SQLite, you'll find that implementation / performance in many databases. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
