24 Mar
2004
24 Mar
'04
10:51 a.m.
Joerg Walter <jhr.walter <at> t-online.de> writes: [...]
(or to define the index as the set of distinct values of the related attribute(s))?
This would call for a new type of index, IHMO. Gotta think it over.
On a second thought: such a beast would be more of a view (losing the relation to indexed_set's original data) and therefore possibly out of scope for your library?
Maybe. Anyway, this is not going to be implemented anytime soon (the roadmap has other top-priority tasks, most notably serialization support and hashed indices.) Joaquín M López Muñoz Telefónica, Investigación y Desarrollo