On Wed, Jun 18, 2008 at 3:15 PM, Jay A. Kreibich <[EMAIL PROTECTED]> wrote: > On Wed, Jun 18, 2008 at 01:39:34PM +0100, Alberto Sim?es scratched on the > wall: >>
[...] > > What is the original ordering of the data? In order to index the > data, it needs to be sorted, and the sort can get expensive due to > page thrashing. In fact it should be random for all columns. >> Can it be caused by a larger db file? > > I suppose, although I'm not sure that's directly related. Index > creation is very sensitive to the number of free pages in the page > cache, but for an index of that size, chances are you're always over. Disk access on that cluster was always a problem. It might be that as well. > Increase the cache size (see PRAGMA commands). Actually, this is > good practice anytime you need to create an index. Thanks Alberto -- Alberto Simões _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
