Prospective agenda for VOQL session in Madrid

[Patricio F. Ortiz]

Hi Yuji,

On Wed, 21 Sep 2005, Yuji SHIRASAKI wrote:
> Hi Patricio,
> 
> > There is however, not a small number of columns (>1800) in the literature
> > with brackets in their names, eg, metal aboundances, forbidden or
> > semiforbidden transitions, etc. How are those cases supposed to be handled?
> 
> I think this hasn't been discussed at the VOQL WG.
> 
> So this is my personal thinking.
> 
> In the SQL standard, a double quatation can be placed by repeating it.
> 
>   e.g.  abc"def --> "abc""def"
> 
> So, if we apply this pattern to the ADQL bracket delimiter,
> 
>   e.g.  [O/Fe] --> [[[O/Fe]]]
> 
> or there are some other proposals ?

The solution above probably makes writing a parser a bit more painful :-)

On the other hand, there are still a number of characters which do not
appear in column names which could be used as delimiters. Looking at
vizier's data (dated last year), the following characters have not been
used in column names:

! @ # $ ^ ~ ; `

If one could say #[O/Fe]# (or any of the other symbols above), it would be
great or perhaps something like ][O/Fe][, ie, a word delimited by a right
bracket-left bracket pair should be left untouched.

Whatever we choose, we should then make it clear what the forbidden
combinations are.

> We need decide how to treat this kind of case.

I fully agree, otherwise the problem will escalate and (I guess)
it wouldn't be too good to have 99%-complaint ADQL parsers around the
world which differ in the treatment (or lack of) "unusual" column names
(or table names as you suggest).

Cheers,

Patricio

---
Patricio F. Ortiz			pfo at star.le.ac.uk
Department of Physics & Astronomy	Phone: +44 (0)116 252 2015
University of Leicester			
Leicester, LE1 7RH, UK