Re: Database naming convention (yet another post of it, but a bit different)

"Pablo Sanchez" <pablo_at_dev.null> wrote in message news:Xns92A2647D86B0Bpingottpingottbah_at_209.189.89.243...
> "Marshall Spight" <mspight_at_dnai.com> wrote in
> news:AJTo9.92754$DN4.13763_at_sccrnsc01:
>
> > "Petruza" <autotacle_at_uol.com.ar> wrote in message
> > news:3e55b51a.0210081559.4ac4c1f9_at_posting.google.com...
> >>
> >> 1) tables names are in singular, not to get confused with adding
> >> or not an ending "S", and because you are refering to a kind of
> >> class from which you would handle single instances.>snipped<

Here's the deal. Celko recommends plural for for table names, and I have used this. Trouble is when going from an ERD ex. EMPLOYEE (works on) zero or many PROJECT, then the entity names resolve in a singuluar table name. But the semantics of the ERD (in this case the cardinality) assume plural. On an ERD this makes sense since the cardinality determines singluar or plural.

    But let's say we always use singular. Do we lose anything? I think not. And if we say Table Employees (plural), does each combination of attribute instances (of course I mean the identifer(s)) describe all employees? Well, of course not.

    So I am back to a table name (or an entity name) that names the class of things described. Singular. Person, Department, etc. And it is perfectly legit to have a table with one tuple. How would you name this? Plural?

    So let's go back with singular Entity and Table names.

Best Regards,

Greg

Greg Received on Sat Oct 26 2002 - 03:46:52 CEST