From: Michael Conroy on
Gentlemen,
I have read all the responses and I thank you for the input. John,
normalizing the data would eliminate null values but my source file is a
little more complicated. I just found out some airports use the GrossIn fuel
amount, some use the NetIn and both fields can have a value so it is not
exactly like a credit debit. I could still normalize it and add a qualifier
that labels it Gross or Net, but I am somewhat leary of modifying source
information if I can use a query to grab what I want. So I will add this
gross/net "switch" to the airport table and use that to grab the correct
column from the fuel table.

Additionally, I just wanted to add that time is a factor in this discussion.
As I mentioned in the original post I am importing a spreadsheet from another
company that tells us how much fuel is pumped into each plane. I realize I
failed to mention that the data is historical and will never be altered. So
seeing a null value does not mean that they have not pumped the fuel yet, it
means they just never put a zero in the spreadsheet cell. Since I am
normalizing other fields before the data gets added to the main table, in
this case, I feel comfortable adding zeros to their source information to
make my life easier down the road.

So with dynamic data, a zero means the tank is empty, null means you did not
look yet. Thanks again and sorry for the late reply.
--
Michael Conroy
Stamford, CT


"John W. Vinson" wrote:

> On Thu, 25 Mar 2010 14:15:01 -0700, Michael Conroy
> <MichaelConroy(a)discussions.microsoft.com> wrote:
>
> >So my question is this, what
> >question should I be asking about the data to determine whether I store a
> >zero or null value in a numeric field?
>
> I'd ask "Does a blank field have a different interpretation than a zero"? In
> some cases it may not - a credit is a good example, where a zero default (and
> nulls being forbidden) makes sense. In other situations they might indeed be
> different - "How many birds are on my lawn" might be zero (I looked and there
> weren't any) or it might be null (I haven't looked and don't have any idea how
> many).
>
> It's got to be decided on the basis of the real-life situation.
> --
>
> John W. Vinson [MVP]
> .
>