MySQL (5.1.41-3ubuntu12.10-log) seems to give predictable results when comparing strings using> (more) and <(less):
select "a" > "a", "a" > "b", "b" > "a", "ab" > "aa", "ab" > "aabbbb"; +-----------+-----------+-----------+-------------+-----------------+ | "a" > "a" | "a" > "b" | "b" > "a" | "ab" > "aa" | "ab" > "aabbbb" | +-----------+-----------+-----------+-------------+-----------------+ | 0 | 0 | 1 | 1 | 1 | +-----------+-----------+-----------+-------------+-----------------+
and also apparently uses keys:
explain select productcode from products where productcode < 'no'; +----+-------------+----------+-------+-----------------+------+---------+------+------+--------------------------+ | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra | +----+-------------+----------+-------+-----------------+------+---------+------+------+--------------------------+ | 1 | SIMPLE | products | range | productcode,ppp | ppp | 34 | NULL | 432 | Using where; Using index | +----+-------------+----------+-------+-----------------+------+---------+------+------+--------------------------+
This doesn't seem to be documented - is it a robust cross-platform feature?
source share