"compound data" - is there a more standard term?

#11

Op 27-4-2010 1:53, Ben Finney schreef:

Quote:

metaperl <metaperl (AT) gmail (DOT) com> writes:

When a database designer / developer packs conceptually separate
values into a single column, what is this called?

Short answer: It violates one of the conditions for first normal form
(1NF). That condition has conflicting definitions, though, so you will
need to be clear on what definition you're referring to.

You will also hear the term “multi-valued” or more formally
URL:http://en.wikipedia.org/wiki/First_normal_form#Repeating_groups>.
This is what many people think of first when asked to come up with an
example of a database that violates 1NF.

Your example, though, is just as good even though it has nothing to do
with repeating groups. So this term is too narrow and doesn't help
describe the practice you're asking about.

That practice violates what E.F. Codd termed 1NF (first normal form),
because it uses values that are not “atomic”. (There are a number of
other conditions for a database to be in 1NF, but this condition of
atomic values is the one relevant to your question.)

So, on that basis, attributes that do not meet this condition could be
called “non-atomic” attributes. The trouble is, “atomic” has no absolute
meaning.

C.J. Date criticises this term for its unclear definition and unclear
application, and suggests the following phrasing of the condition as a
way of avoiding the ambiguity of “atomic”:

“Every row-and-column intersection contains exactly one value from
the applicable domain (and nothing else).”

He further criticises the condition itself as Codd defined it, for
excluding tuple-valued attributes and relation-valued attributes, which
he considers important and valid in the relational model.

So, on that basis, attributes that do not meet this condition could be
called “domain-violating” attributes. The trouble is, “domain” can be
defined by the DBA to mean whatever they like and allow any arbitrary
values for a particular attribute.

For more, see the Wikipedia page for 1NF, especially the discussion of
URL:http://en.wikipedia.org/wiki/First_normal_form#Atomicity>.

Luckily i'm not bound to 1NF...., and i will take the consequences when
i make a wrong judgement about violating these 1NF's...

;-)

--
Luuk

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

"compound data" - is there a more standard term?

comp.databases comp.databases

Re: "compound data" - is there a more standard term? - 04-27-2010 , 01:40 PM