Path: dp-news.maxwell.syr.edu!spool.maxwell.syr.edu!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!postnews.google.com!h48g2000cwc.googlegroups.com!not-for-mail
From: "dawn" <dawnwolthuis@gmail.com>
Newsgroups: comp.databases.theory
Subject: Re: Columns without names
Date: 16 Sep 2006 08:44:59 -0700
Organization: http://groups.google.com
Lines: 45
Message-ID: <1158421499.274645.83490@h48g2000cwc.googlegroups.com>
References: <1158372047.852132.130270@p79g2000cwp.googlegroups.com>
NNTP-Posting-Host: 207.199.203.250
Mime-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
X-Trace: posting.google.com 1158421504 3286 127.0.0.1 (16 Sep 2006 15:45:04 GMT)
X-Complaints-To: groups-abuse@google.com
NNTP-Posting-Date: Sat, 16 Sep 2006 15:45:04 +0000 (UTC)
In-Reply-To: <1158372047.852132.130270@p79g2000cwp.googlegroups.com>
User-Agent: G2/1.0
X-HTTP-UserAgent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7,gzip(gfe),gzip(gfe)
Complaints-To: groups-abuse@google.com
Injection-Info: h48g2000cwc.googlegroups.com; posting-host=207.199.203.250;
   posting-account=fLxfag0AAADzmFPc_IBdM-stGY4WCEYa
Xref: dp-news.maxwell.syr.edu comp.databases.theory:44926


JOG wrote:
> Faced with a data collection something like:
> Tom is aged 20, Dick is aged 30 and Harry is aged 40
>
> I find it apt to view a relation predicate for them as:
> "There is a people_relationship where name is X and age is Y"
> (Initially this was to stop any urge to think in terms of entities as
> opposed to assertions, and even though this is now unnecessary the
> process has stuck)
>
> However it struck me that this process may be reversed. Not very
> interesting, until I started considering domain-defining statements
> such as:
>
> "There is a number, 7." or "There is a letter, b."
>
> These are potentially different statements to those such as "There is a
> person, Sally" because the latter is stating "There is a person where
> name:Sally". The values held in the former assertions have no
> discernable attribute name - rather than a tuple such as {
> (name:Sally), (age:28)} I just have { (7) } or { (b) }.
>
> So I would like to offer for discussion the concept of whether it is
> possible to have a relation with a single column /but no column name/.
> Granted it is kooky, but is there anything theoretically against this
> principle, and if not, could it be of value?

Since we define what a "database relation" is, we could define it to
permit a column with an empty string as the name, although it could
surely be argued that is not "no name" but a name that is the empty
string.  If the rules of the game include that columns must be uniquely
named within a relation, then only one column per relation would be
permitted to use the empty string.

Then, of course, you need a tool that implements relations in a way
that permits this. Some of the tools I use do not require that a column
have a name, but then you refer to it by positioning.  That is
typically considered a no-no in database relational theory (although
not in mathematical relations), but one could certainly define a
relation that way too, as these tools do.  You can then have any number
of attributes with no column name.

Cheers!  --dawn