Sorry, hit 'send' too early.
On 22 Aug 2013, at 6:15pm, Igor Tandetnik <[email protected]> wrote:
> But again, by what formal mechanism does a property of the column affect the
> behavior of the operator?
I see no reason for it to do that. I can see why the property of the column
might affect the behaviour of an index made on that column but there are
problems if it affects comparisons made using values from that column. For
example, the meaning of something like
create table t(x text collate nocase,
y text collate rtrim);
select * from t where x = y;
is not at all clear purely from reading the SQL code. You have to have
actually read the picky details. And the consequences of the picky details are
that under SQLite3, these two queries:
select * from t where x = y;
select * from t where y = x;
can return different results. Which I find annoying.
But my original problem, as stated, isn't with this. It's just with the way
that one can explicitly apply a correlation in an expression by using the
COLLATE operator. I feel that it doesn't reflect the real-world operation of
correlations by SQLite.
Simon.
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
