On 8/22/2013 8:52 AM, Simon Slavin wrote:
Nevertheless do you understand the point I'm trying to make -- that collations
are a modifier for comparisons not individual values ?
I do understand your point. I still don't understand how it's supposed
to apply in practice to a situation like
create table t(x text collate nocase);
select * from t where x = 'a';
Here, at the time I specify "collate nocase", there is no comparison it
could apply to. So what is it a property of, in your thinking? By what
mechanism does it end up applying to x='a' comparison (I assume to do
want the statement to return rows both with 'a' and 'A' in column x)?
How would you modify the formal spec at http://sqlite.org/datatype3.html
to lead to your desired outcome? Precise wording matters.
As I said, I could relate to your point of view better if you just did
this: in http://sqlite.org/datatype3.html section 6.1, replaced two
occurrences of "with precedence to the left operand" with "It's an error
if two operands have different collations". This keeps the existing,
well defined mechanisms intact, while neatly excluding the case you seem
to find most objectionable.
--
Igor Tandetnik
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
