Re: Counting propositions
Date: 18 Jun 2004 12:17:47 -0700
"x" <x-false_at_yahoo.com> wrote in message news:<40d2cceb$1_at_post.usenet.com>...
> **** Post for FREE via your newsreader at post.usenet.com ****
> "Tony" <andrewst_at_onetel.net.uk> wrote in message
> > > You see, there might be a relational database that has no named
> > > relation/relvar that (directly) assert employment.
> > A badly designed, and probably un-normalised, database I would suggest
> > ;-)
> Well, you cannot always choose the design :-)
... in which case, maybe you can't get useful information from it.
> Why unnormalized ?
Because you apparently have no table in which emp_id is unique, and so nowhere for a fact about an employee to be stored once and only once.
> > > It works only when you count Emp_IDs, not the propositions employing
> > > :-)
> > No, it always works when you count propositions whose intention is to
> > state the existence of an employee. Of course it doesn't work if you
> > count propositions about project assignment or the employees' CD
> > collections etc.!
> In order to count propositions whose intention is to state the existence of
> an employee, you rely on Emp_IDs one-to-one corespondence with the
> employees, so you really count those Emp_IDs. :-)
You certainly can count emp_ids. Or you can count whole propositions, if you know that those propositions have a one-to-one correspondence.
> > But, for every table, "select count(*)" provides useful information -
> > provided you understand what the table means.
> I just showed you that you don't need to count the propositions for counting
> the employees. :-)
> You only need to count the Emp_IDs ... :-)
... the emp_ids in that table yes - which amounts to the same thing, by definition. It would be no good counting the emp_ids in a table that didn't have propositions about EVERY employee - e.g. if not all employees are assigned to projects, then counting the emp_ids in the assignment table would tell you how many employees are assigned to projects, not how many employees there are period. Received on Fri Jun 18 2004 - 21:17:47 CEST