Joe Celko wrote:
>encoding of the set structure, which you can see by the fact that it
>introduces the rgt and lft columns which are what? <<
>Yes, that's right! And numerals are not really numbers, but only an
>ASCII encoding of them. And employee ids are not really employees.
>That is what a model is all about -- representing a fact in symbols.
You've spent a lot of finger energy in these newsgroups arguing
that data should be verifiable. But you've never answered the
question of how the values of lft and rgt are verifiable from the
real world. How do you look at an employee and deduce his or
her lft or rgt attribute? The question is a good one. rgt and lft are
They seem completely artificial to me (which I don't find problematic,
but you seem to).
So long as no one has ever been fired.
And what kind of computer do you have that doesn't do things
sequentially? Even with multiple processors, if the table exists on
a single device, everything happens sequentially.
Um, don't you need to exclude rows where boss_emp_nbr IS NULL from your
Isn't check (select count(*) from OrgChart where boss_emp_nbr is null) = 1)
a little simpler way to check "only one NULL in the boss_emp_nbr column" ?
-- Steve Kass
-- Drew University
-- Ref: 6E27E3DB-F383-4D3C-8377-3C2CBEEC8BBA