"Kerry, Richard" <richard.kerry at siemens.com> writes: > The hash is not expected to be unique, it just provides a starting point > for another search (usually linear ?). The database is good at organizing indexes and searching in them. Why not let the database do what it's good at.