[TAP] case sensitivity of query

[Francois Ochsenbein]

Just to emphasize on Tom's comments about SQL case sensitivity:

A + B. the language is defined as case insensitive ;
       ADQL recommendation clearly specifies case 
       insensitivity (section 2.1.2).

C. table + column names: ADQL recommendation specifies
   "ADQL allows making use of the quoting mechanism
   to handle the case sensitiveness if needed"
   A good practice would be, in my opinion, to
   use the case of the catalog/schema/table/column
   names as those are spelled in the VOResource or
   in the TAP_SCHEMA (hopefully these 2 should be
   identical!)

D. character variable as data: upper and lower case
   differ, and there are also problems with the
   latin extensions or the other alphabets. I didn't
   see anything about ISO-8859 or other codes in SQL92
   (not talking of unicode which did not exist :-)
   It seems that SQL defines the "upper" and "lower" 
   functions (not mentioned in the ADQL document,
   but listed as 'reserved words'), i.e. a caseless
   equality is written as    
       upper(t1.col1)=upper(t2.col2)
  
About this last point (case sensitivity of strings)
postgres provides the 'ilike' operator (case-insensitive
like operator), and allows also the creation of
caseless indices (faster searches with ilike operator)
But it's not SQL-92 standard.

This brings up another motivation to add some definition of
the column domains into TAP_SCHEMA (only uppercase or only
lowercase strings represent a domain definition, as well
as restriction to 7-bit ascii characters).

Francois
=======================================================================
Francois Ochsenbein    ------   Observatoire Astronomique de Strasbourg
   11, rue de l'Universite 67000 STRASBOURG  Phone: +33-(0)390 24 24 29
Email: francois at astro.u-strasbg.fr (France)    Fax: +33-(0)390 24 24 17
=======================================================================