Re: object naming conventions

From: Paul Brewer <paul_at_paul.brewers.org.uk>
Date: Tue, 1 Jul 2003 20:59:17 +0100
Message-ID: <3f01f5fa_1@mk-nntp-1.news.uk.worldonline.com>

 "Daniel Morgan" <damorgan_at_exxesolutions.com> wrote in message news:3F01D6F4.A7E79DBC_at_exxesolutions.com...
>
> I find naming conventions meaningless for tables other than to separate
> them from views and materialized views. Where I get concerned is with
> constraint names as it is near meaningless to report a constraint
> violation of SYS_00.....
>

Agree completely. I hate system generated constraint names. Smacks of sheer laziness, or ignorance.
It screws up schema compares too, with most tools. To digress slightly, along the (very sensible IMHO) lines of trying to get meaningful error messages from the database itself, rather than from some 'wrapper' layer, one particular feature has irritated me for years:

Database:
'Inserted value too large for column'
Paul (from the rooftops, in a voice to wake the dead): 'WHICH BLOODY COLUMN?' Regards,
Paul

P.S. To OP, by the way: This is purely personal preference but I like:

Meaningful non-prefixed singular names for tables and columns
Meaningful non-prefixed plural names for views
Meaningful non-prefixed names for functions ('TO_BASE(orig_curr, amount)',

for example, makes the SQL read more easily)

Prefixes for most other objects; PK_, FK_, CK_ for constraints

SQ_ for sequences
SP_ for stored procedures
PG_ for packages - but I make exceptions for common ones like 'DEBUG' as in

(made up example)

when others then

    debug.send('Error in PG_SOMETHING: '||v_ora_error||' in section '||v_section||' while processing Customer No. '||to_char(l_customer_id));

Regards,
Paul

Regards,
Paul Received on Tue Jul 01 2003 - 14:59:17 CDT