On Sun, 31 Jan 2016 18:02:38 +0100
Tom Lane <tgl@xxxxxxxxxxxxx> wrote:
> Harald Fuchs <hari.fuchs@xxxxxxxxx> writes:
> > Ben Leslie <benno@xxxxxxxxxxx> writes:
> >> "Technically, PRIMARY KEY is merely a combination of UNIQUE and NOT NULL"
> >>
> >> I wanted to clarify if that was, technically, true.
>
> > Yes, but see below.
>
> >> "identifying a set of columns as primary key also provides metadata
> >> about the design of the schema, as a primary key implies that other
> >> tables can rely on this set of columns as a unique identifier for
> >> rows."
>
> Yeah. The extra metadata has several other effects. Perhaps it would be
> better to reword this sentence to make it clear that PRIMARY KEY is
> equivalent to UNIQUE+NOTNULL in terms of the data constraint that it
> enforces, without implying that there is no other difference. I'm not
> sure about a short and clear _expression_ of that though ...
How about:
"PRIMARY KEY is merly a combination of UNIQUE and NOT NULL with regard
to data consistency behavior."
"identifying a set of columns as primary key also provides metadata about
the design of the schema, as a primary key implies that other tables can
rely on this set of columns as a unique identifier for rows. This
metadata may be used by external programs, but is also utilized interally
by the server in some cases."
Do we have to be so vague...
A PRIMARY KEY enforces a UNIQUE, NOT NULL constraint and additionally allows the server to exhibit additional behaviors based upon the additional knowledge that the chosen constraint unique identifies a specific record. The behaviors are: <list them here>. External programs and extensions may also make use of the additional meta-data communicated through the use of PRIMARY KEY instead of a simple UNIQUE+NOTNULL constraint.
David J.
