On Fri, 12 Jan 2001 05:56:43 Oliver Elphick wrote:
>Rajit Singh wrote:
>>I have five tables which are really intended to be one big table, but
>>has been adapted to comply with Postgres's (irritating) 8kb tuple size
>>limit (ver 7.0.3).
>>
>>I have 'modtime' fields in each of these tables which record the last
>>time a record was updated. But I would like the behaviour to be such
>>that a modtime field in each of these associated tables is updated when
>>any of the tables are updated. And I would like all of the modtime
>>fields to take the same value if possible.
>
>If the rows in these five tables are meant to be all one, why have five
>separate modtime fields? That breaks the "rules" against redundant data.
>Just have it in one table and put triggers to set it in all five.
Why not just have one modtime field in the main table and just include the
primary key in the other tables? It sounds like you're going to include
records from each table based on the primary key anyway. There is no need to
duplicate the modtime field across the other table if the columns in these
tables all actually represent a single record.
--
Anthony E. Greene <agreene@pobox.com> <http://www.pobox.com/~agreene/>
PGP Key: 0x6C94239D/7B3D BD7D 7D91 1B44 BA26 C484 A42A 60DD 6C94 239D
Chat: AOL/Yahoo: TonyG05 ICQ: 91183266
Linux. The choice of a GNU Generation. <http://www.linux.org/>