Stephan Szabo <sszabo@megazone.bigpanda.com> writes:
> Per the report from Clark C Evans a while back and associated discussion,
> it seems like recent versions of the SQL spec changed the rules for
> foreign key column references such that the columns of the referenced
> unique constraint must be named in order (this is somewhat silly since
> unique(a,b) really should imply unique(b,a) but...).
I do not believe that that reading is correct. If the SQL committee had
intended such a change, it would surely have been called out as a
compatibility issue in Annex E of SQL2003. Which it isn't.
What I see in SQL99 is (11.8 <referential constraint definition>)
If the <referenced table and columns> specifies a <reference column list>, then the set of <column name>s
contained in that <reference column list> shall be equal to the set of <column name>s contained in the <unique
column list> of a unique constraint of the referenced table. Let referenced columns be the column or columns
identifiedby that <reference column list> and let referenced column be one such column. Each referenced column
shallidentify a column of the referenced table and the same column shall not be identified more than once.
where SQL2003 has
If the <referenced table and columns> specifies a <reference column list>, then there shall be a one-to-one
correspondencebetween the set of <column name>s contained in that <reference column list> and the set of <column
name>scontained in the <unique column list> of a unique constraint of the referenced table such that corresponding
<columnname>s are equivalent. Let referenced columns be the column or columns identified by that <reference column
list>and let referenced column be one such column. Each referenced column shall identify a column of the referenced
tableand the same column shall not be identified more than once.
I think SQL2003 is actually just trying to say the same thing in more
precise language: you have to be able to match up the columns in the
<reference list> with some unique constraint. I don't think the "one
to one" bit is meant to imply a left-to-right-ordered correspondence;
that's certainly not the mathematical meaning of a one-to-one function
for instance.
> The information_schema definition seems to require this in order for
> one to use the information to find out the column references.
I'm more inclined to think that we've messed up the information_schema
somehow ...
regards, tom lane