DB schema -> graph, good strategies

Howdy!

I'm working on doing keyword searching in relational databases. For this purpose it is not uncommon to build database-wide indices, which means that locating the keywords themselves becomes pretty fast (certainly beats full-table scans on all relations).

Now, the database schema is often used as a basis for building a graph which is used in many very different ways to determine relevancy or rank results. The papers I've read are at best superficial on why they use the strategy that they do for building schema graphs. If primary/foreign key relationships map directly to edges, in which direction should they go, if they're directed at all? Can certain 'patterns' in the graph be exploited (eg. to find relationships that are 'lost' in normalisation etc.)? These are the kind of issues I'd like to explore, but I've been able to find only very little literature to start from. Any pointers to literature or discussion would be gratefully welcomed.

Martin

• -- Homepage: http://www.cs.auc.dk/~factotum/ GPG public key: http://www.cs.auc.dk/~factotum/gpgkey.txt
  -----BEGIN PGP SIGNATURE-----
  Version: GnuPG v1.0.6 (GNU/Linux) Comment: Using Mailcrypt+GnuPG <http://www.gnupg.org>

iEYEARECAAYFAj25kDEACgkQYu1fMmOQldWxoQCeJIZHwzPARJbnHS4gMicFFFcu koQAn3MrA7TbwVmwxYjLkTPl/0I7GF17
=YS3F
-----END PGP SIGNATURE-----
Received on Fri Oct 25 2002 - 20:40:49 CEST