Subject | Re: [ib-support] Re: Please remove the ambiguity check |
---|---|
Author | Ivan Prenosil |
Post date | 2001-12-07T15:35:17Z |
IF there were clearly defined behaviour (e.g. get ambiguous field from last table it is defined in),
then PERHAPS one could think about whether ambiguity tolerate or not.
But as Ann said, IB/FB is NOT designed that way. It is up to optimizer which
field will be chosen.
Ambiguous queries are BAD the same way as using non initialized variables -
it is only good luck if the program works as expected, and you are never sure
when it stops working.
(And IMO it would be just wasting of time to implement detection
of special cases where ambiguous fields are intechangeable because
they contain the same value.)
What about implementing following extension
SELECT INTEGER FROM tab;
it will fetch arbitrary (but integer only) column from tab. Nice ? :-)
Ivan
http://www.volny.cz/iprenosil/interbase
then PERHAPS one could think about whether ambiguity tolerate or not.
But as Ann said, IB/FB is NOT designed that way. It is up to optimizer which
field will be chosen.
Ambiguous queries are BAD the same way as using non initialized variables -
it is only good luck if the program works as expected, and you are never sure
when it stops working.
(And IMO it would be just wasting of time to implement detection
of special cases where ambiguous fields are intechangeable because
they contain the same value.)
What about implementing following extension
SELECT INTEGER FROM tab;
it will fetch arbitrary (but integer only) column from tab. Nice ? :-)
Ivan
http://www.volny.cz/iprenosil/interbase