Re: Relation Definition
Date: Tue, 22 Feb 2005 02:14:53 GMT
Message-ID: <xAwSd.3039$873.584_at_newsread3.news.pas.earthlink.net>
>> No, that cannot be right because it would allow the same attribute name
>> to appear more than once.
Why so? The header is a set -- so the elements i.e. the named typed attributes in a set are unique. When one ignores the typed perspective, the header is simply a set of column names, again being a set, the elements must be distinct.
If the uniqueness of attribute names were not made explicit is your concern,
the definition of the header in a relation is stated as: Given a collection
of n types or domains Ti ( I = 1, 2, ...n ), not necessarily distinct, r is
a relation on those types if it consists of two parts, a heading and a body,
where the heading is a set of n attributes of the form Ai:Ti, where Ai (
which must all be distinct) are the attributes names of r and the Ti are
corresponding type names ( i = 1,2,...n )
>> http://mathworld.wolfram.com/n-Tuple.html
That provides a possible vector interpretation of an n-Tuple. Nothing in
that reference suggests the notion of n-dimensional tuple is usually
-- AnithReceived on Tue Feb 22 2005 - 03:14:53 CET