Archives of the TeradataForum

Message Posted: Mon, 19 May 2008 @ 21:20:45 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

First, I would discourage using upper vs. lower to test for numeric. Upper('!#$') = lower('!#$'), but that's not numeric.

Now, about the results you're seeing. Consider this:

     select type('1234');

     VARCHAR(4) CHARACTER SET UNICODE

     select type( lower('1234') );

     VARCHAR(4)

     select type( upper('1234') );

     VARCHAR(4) CHARACTER SET UNICODE

This explains why the char2hexint results are different, but it seems odd that lower('1234') would be Latin when the input string is Unicode. Nevertheless, they will compare as equal.

There is no built-in function to test for numeric; you must provide your own or use one of the methods suggested by others.

	<Prev	Next>		<<First	<Prev	Next>	Last>>