It's clear that you can use numeric characters in SQL table names and use them so long as they're not at the beginning. (There's a discussion here on one of the side effects: SQLite issue with Table Names using numbers?) The database I'm targetting is Oracle 10g/11g.
I'm designing a reporting database where naming some of the entities clearly is best done by describing the reports, which are named after numbers ('part 45', '102S', '401'). It's just the business domain language: these reports just aren't commonly referred to by any other name. The entities I'm modelling really are best named this way.
My question is: am I going to have difficulties with maintenance or programmability if I put numbers in a table name? I'm always worried about ancillary software around the database: drivers, ETL code that might not play nice with a non-plain-vanilla name. But there's a real benefit in intelligibility in this business domain, so am I just being squeamish?
My question put simply is: are there any 'gotchas' or corner cases that would rule out a table name like PART_45_AUDIT?