Table names, VOSI vs. TAP_SCHEMA

Mark Taylor m.b.taylor at bristol.ac.uk
Tue May 5 13:04:03 CEST 2015


On Mon, 4 May 2015, Markus Demleitner wrote:

> Gilles,
> 
> On Mon, May 04, 2015 at 11:01:03AM +0200, gilles landais wrote:
> > However, this ambiguity could be avoided with removing the schema name of
> > the table_name (my implementation which concatenate the schema_name and the
> > table_name  was not a good idea).
> 
> Well, it was not a good idea because what came out is not a valid
> ADQL table reference.  Other than that, I think the TAP author's
> solution to the underlying problem (enumerate all tables that can be
> referenced in queries to this service) was sound, and splitting
> things up is asking for trouble in the presence of DBs running
> primarily on catalogs rather than schemas and other DBs that do
> namespacing tricks to hide schema names from users.
> 
> Hence I believe that VODataservice -- your description:
> 
> > The XML tapschema  describes the schema in a XML node containing the schema
> > name and the tables of the schema. So it is not needed to give the schema
> > name again in the table_name.
> > 
> > <schema>
> > <name>viz4</name>
> >      ....
> > <table type="base_table">
> > <name>B/avo.rad/catalog</name>
> > ....
> >      </table>
> > ....
> > </schema>
> 
> -- is asking for trouble.  I (and, for dramatic effect, I'm now
> putting on my Registry chair hat) believe the VODataService must be
> fixed here, both for proper support of delimited identifiers and for
> effortless referencing of tables regardless of naming schemes.

Gilles' example above is only troublesome because of the missing quotes.
As TAPVizier is currently set up, it ought to read:

   <schema>
   <name>viz4</name>
        ....
   <table type="base_table">
   <name>"B/avo.rad/catalog"</name>
   ....
        </table>
   ....
   </schema>

This is OK because in TAPVizier you do not need to qualify tables
by schema in queries ('SELECT * from "B/avo.rad/catalog"' works).

Mark

--
Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
m.b.taylor at bris.ac.uk +44-117-9288776  http://www.star.bris.ac.uk/~mbt/


More information about the dal mailing list