VODataService becoming WD, then PR
patrick.dowler at nrc-cnrc.gc.ca
Mon Nov 17 11:18:52 PST 2008
On 2008-11-17 09:58:55 Ray Plante wrote:
> I'll suggest that the simpler the solution, the better it stands up to the
> "this needs to be prototyped" argument. By prototype, I think this means
> that we can point to an actual use of this by an application. The
> extension mechanisms hopefully provide a solution to the chicken-and-egg
> One minor comment. Table names are required to be unique within an entire
> tableset and already include any schema (or catalog) names. Thus,
> <targetSchema> would be superfluous.
> more comment out there?
Sure, I will also add that I think this metadata is vital to using real
databases. I agree that if prototyping is required that it is on the client
side since just implementing some metadata exposure in a service does not
However, this particular problem is well understood and relatively simple,
which makes me think one can prove analytically that a proposed solution
is "correct" or "sufficient". You just have to look at the syntax of the JOIN
clause to see what is needed: a pair of tables and one or more pairs of
columns that can be compared for equality.
From Gerard's idea, isn't it missing the other table? That is, should it be:
<xs:element name="fromTable" type="xs:string"/>
<xs:element name="targetTable" type="xs:string"/>
<xs:element name="fkColumn" type="vs:FKColumn" maxOccurs="unbounded"/>
(since Ray already pointed out that schema is embedded in the table name).
Presumably fkColumn needs minOccurs="1", but maybe that is the default.
Tel/Tél: (250) 363-6914 | fax/télécopieur: (250) 363-0045
Canadian Astronomy Data Centre | Centre canadien de donnees astronomiques
National Research Council Canada | Conseil national de recherches Canada
Government of Canada | Gouvernement du Canada
5071 West Saanich Road | 5071, chemin West Saanich
Victoria, BC | Victoria (C.-B.)
More information about the registry