Set Like Data

I am involved in a project for which Postgresql is the
preferred DBMS.  I am experienced with many DBMS
products, but am new to Postgresql.  The project
involves many cases of a particular set-like structure
that I am unsure how to represent in Postgresql.

If I may illustrate:

Assume two tables, TABLE_A and TABLE_B, each of which
contains an attribute att_A and att_B which may have
multiple values in each case, each of which may be a
key to a record in TABLE_C.  Thus, there are
many-to-many relationships between TABLE_A or TABLE_B
on the one hand, and TABLE_C on the other.  The
obvious way to encode these relationships in
Postgresql is with two join files.

However, there are many set-like questions that occur
in the problem domain that all have the same basic
structure.  I cannot see how to code this in a
reasonably efficient manner:

Given a specific record in TABLE_B, there will be a
set of keys from TABLE_C such that there are joins
between the record in TABLE_B and records in TABLE_C.
For each record in TABLE_A there will a similar join
set.  We wish to determine the set of records in
TABLE_A such that its join set is a subset of the join
set between the record in TABLE_B and TABLE_C.

Just a note in the environment: RedHAt Linux, Python
interface to Postgresql v. 7.0.2.

Perhaps there is something better than join files for
this that I have not discovered yet.  Does anyone have
any helpful comments on how to imlement this in a
reasonably efficient manner?

Regards, and thanks in advance
John Bell


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.
http://personal.mail.yahoo.com/