In a message dated 4/4/2004 11:30:24 PM Eastern Standard Time, [EMAIL PROTECTED] writes:
> >Personally I would be surprised if either database had a > way of handling > >leading wildcards other than an exhaustive scan. > > Use "Contains"/English Query. See MS-SQL Server Docs. > > Joe Eugene Joe you didn't understand my point. What I was saying wasn't whether you could FORM a query to do this, but rather exactly HOW the query would be executed (in the run engine). What I was saying is a query on trailing wildcards does not need to do an exhaustive scan, since BRAD PITT doesn't match SARA% and you can tell right away that it doesn't, you don't have to read the whole entry to determine that. However to determine if ABERCROMBIE matches %CROMBIE you DO have to read the entire entry. And in fact you have to read every index entry. Whereas with SARA% you only need to read those entries that start with SARA% and in a B-Tree these are clumped together, not spread out all over the tree. With leading wildcards, the index entries that may match could be spread out all over the tree nodes, there's no way to tell. -Will -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users
