Is there a term in database theory for the principle of breaking the
data down into its smallest constituent parts rather than putting data
that might need to be processed separately in the same field?

The example that is usually used is names. When storing people's names
in a database, it is generally better to make separate fields for
First_Name, Last_Name, Middle_Name, etc., rather than just having one
Name field.

Is there a technical term for that?

Is it part of one of the normalization levels?

On Mon, 21 Jun 2010 18:47:41 -0700, Prof Wonmug <wonmug@e.mcc> wrote:

Normalisation?

I think so.

There was a genealogy program, called Personal Ancestral File, which had
fields for four names, one of which was the surname. But what you entered in
those fields was not stored in those fields. What was sotred was a pinter to
the names database, so that, for example, the name "Peter" was stored only
once in the database. And when it appeared in any of the name fields, there
would be a pointer to it from that field.

That is part of the normalisation principle of having no redundant data.


--
Steve Hayes from Tshwane, South Africa
Web: http://hayesfam.bravehost.com/stevesig.htm
Blog: http://methodius.blogspot.com
E-mail - see web page, or parse: shayes at dunelm full stop org full stop uk

Prof Wonmug <wonmug@e.mcc> writes:

I believe it's called “functional decomposition� and it's not specific
to databases <URL:http://en.wikipedia.org/wiki/Functional_decomposition>.

Note that important preconditions for applying functional decomposition
are that the compound value always has unambiguous component parts, and
that the re-composition from those parts can be done mechanically
whenever needed.

So, for instance, a timestamp (date-and-time) value is a good candidate
for functional decomposition by the above preconditions.

Oooh, no. That's a very common trap. Consider the names for a Spanish
person, a Dutch person, a Chinese person, a Ukranian person, a Korean
person:

* What are the “first name�, “middle name�, and “last name� components
of the name?

* What is a typical query going to expect will go into each of those
components to match a given person? Does it correspond unambiguously
with the criteria used when someone else first decomposed the name?

* Once you have the component values decomposed, what is the single
mechanical process of re-composition to get the full name?

It's not at all clear that people's names are a good target for
functional decomposition.

A better example is street addresses (though, of course, there are some
locations that make this not as simple as it might first appear).
Country, state or province, postal code, city or town, street, property
number, unit number, addressee's full name. That's much less ambiguous
and more reliably queried and re-composed.

--
\ “The power of accurate observation is frequently called |
`\ cynicism by those who don't have it.� —George Bernard Shaw |
_o__) |
Ben Finney

On 2010-06-22, Ben Finney <bignose+hates-spam (AT) benfinney (DOT) id.au> wrote:
Not really. There are national variations in addressing too,






--- news://freenews.netfront.net/ - complaints: news (AT) netfront (DOT) net ---

On Tue, 22 Jun 2010 05:04:13 +0200, Steve Hayes
<hayesstw (AT) telkomsa (DOT) net> wrote:

Yes, but this is not what I am asking about.

My question has nothing to do with redundancy. It's a completely
different concept and I was wondering if there is a specific term for
it.

On Tue, 22 Jun 2010 14:18:47 +1000, Ben Finney
<bignose+hates-spam (AT) benfinney (DOT) id.au> wrote:

Hmmm... I hadn't heard that term before. It certainly describes the
principle. I just thought I had heard a different term, but it was
very long ago in school.

The link in that article to "database normalization" does not contain
that phrase.

Thanks

Jasen Betts <jasen (AT) xnet (DOT) co.nz> writes:

Yes, and if you read my message again you'll see that I allowed that. I
maintain that, compared to people's names, addresses are much less
ambiguous in their components and much more easily re-composed from
components by simple mechanical process, and thus a better candidate for
functional decomposition.

--
\ “Reality must take precedence over public relations, for nature |
`\ cannot be fooled.� —Richard P. Feynman |
_o__) |
Ben Finney

Prof Wonmug <wonmug@e.mcc> writes:

I don't think it's an important part of database normalisation; rather,
it's a more general programming technique.

--
\ “The restriction of knowledge to an elite group destroys the |
`\ spirit of society and leads to its intellectual |
_o__) impoverishment.� —Albert Einstein |
Ben Finney

In article <0pe126ts5ep2k7i79snsakvaudr0aidve6 (AT) 4ax (DOT) com>, wonmug@e.mcc
says...
Normalization and this breaking down of fields into component parts (I
hadn't heard the term "functional decomposition" before) aren't the same
thing, though they are related. WRT the street address thing, I
personally think breaking down the street address into multiple pieces
is a mistake, but the city, state and postal code should be their own
fields.

In article <87k4prnq1c.fsf (AT) benfinney (DOT) id.au>, bignose+hates-
spam (AT) benfinney (DOT) id.au says...
I disagree. Basically every culture has "surname" and "given" personal
name components (or in some cases, just a single name). It's just the
order that they are written (or spoken) will vary. Addresses have a
much wider range from what I've seen.

D