From: Kalle Olavi Niemitalo on
--CELKO-- <jcelko212(a)earthlink.net> writes:

> COALESCE (NULLIF (<exp>, <exp>, 1), 0)

MSDN says NULLIF takes only two arguments.
http://msdn.microsoft.com/en-us/library/ms177562(SQL.105).aspx

>> COALESCE (<exp> * 0 + 1, 0)
>
> .. and it does not work with temporal, approximate numeric and string
> expressions.

With "approximate numeric", do you mean floating-point infinities
and NaN values? If not, I don't see what would go wrong, even
with complex numbers or intervals; multiplying any of those with
zero should result in a zero of the same type.

> Do you remember Algol 60?

Unfortunately not.
From: Tony Rogerson on
> No, attribute and tuple are from the relational model or relational
> data base theory. Date has a lot to say on (tuple/row/record),
> (attribute/column/field) and (relation and relvar/table/file)
> differences.

That is correct Attribute and Tuple are indeed from Relational Database
theory which is what I said....

Those terms are not SQL terms - you said RDBMS and not SQL.

Remember - you said "Columns are not fields -- you are still thinking in
file system terms and not RDBMS".

I'm fully aware of what Date says - I've read it - thank you; now - have
you?

--ROGGIE--

"--CELKO--" <jcelko212(a)earthlink.net> wrote in message
news:87e28df4-ddbb-4666-a0e8-deff67a65498(a)x23g2000prd.googlegroups.com...
>>> You are using the term RDBMS, in a Relational Database Management System
>>> we use Attributes and Tuples. <<
>
> No, attribute and tuple are from the relational model or relational
> data base theory. Date has a lot to say on (tuple/row/record),
> (attribute/column/field) and (relation and relvar/table/file)
> differences.