Hello! On Sunday 20 September 2009 08:05:04 Darren Duncan wrote: > The more general solution here to the duplicate column name problem is to be > stricter than the SQL standard and treat attempts to return duplicate column > names as a syntax or constraint error. For example, if you had 2 tables > 'foo' > and 'bar' with columns named (a, b) and (b, c), then a plain "select * from > foo > inner join bar on ..." should throw an exception because there would be two > 'b' > in the result. And so, proper NATURAL or USING behavior is one way to say > "select * from foo inner join bar ..." with success, and spelling out the > result > column list rather than using "*" is another way. But you have to deal with > it > explicitly or the SQL will refuse to run, is what the DBMS should do, or the > DBMS should be customizable so it can be thusly strict.
It's interesting. The new pragma "unique_column_names" may be helpful for a lot of situations same as the "indexed by" condition for selects. Best regards, Alexey Pechnikov. http://pechnikov.tel/ _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users