Re: Why does 1:3 relationsihp require another table?

Tech-Archive recommends: Fix windows errors by optimizing your registry

From: Steve Kass (skass_at_drew.edu)
Date: 05/15/04

  • Next message: Prabhat: "Re: Char and Varchar" 
    Date: Sat, 15 May 2004 02:57:53 -0400

    Anith Sen wrote:

    >Steve,
    >
    >
    >
    >>>doesNotExist is what some people call NULL when they extend non-NULL
    >>>
    >>>
    >relational algebra and step into the muddy waters of 3VL. The fact that I
    >didn't call it NULL is certainly irrelevant. <<
    >
    >I think, in the formal sense, it is relevent. By not calling it as NULL, one
    >can perhaps consider doesNotExist and 123 to be values from the same domain
    >and thus can become vital to the definition of 1NF.
    >
    >
    Thanks, Anith.

    I'm not sure what you mean when you say that what I call it is relevant,
    but I was trying (and didn't do so well) to point out that NULL is
    simply an additional symbol, and one can build a theory like relational
    algebra based on tuples that can contain that symbol in place of any
    domain value. I didn't mean to suggest that I would use a "I'm sure
    it's never going to be a password" flag-like symbol there.

    SK

    >>>SQL is an attempt to implement relational algebra with NULLs that mean
    >>>
    >>>
    >UNKNOWN.<<
    >
    >NULL in SQL does not mean UNKNOWN; in fact it has has no inherent meaning.
    >Instead, people assign arbitrary meanings based on what the lack of a value
    >signifies at a given context.
    >
    >
    Correct, but is it fair to say the development of SQL was guided by a
    desire to make NULL a good representation of a real-world unknown more
    than it was guided by a desire to make NULL a good representation of
    NoValueHere?

  • Next message: Prabhat: "Re: Char and Varchar"