At 10:43 AM -0500 1/17/08, Fowler, Jeff wrote:
Hello All, I've used SQL Server for over 15 years, Oracle off & on when I have no choice, but SQLite for a couple weeks. I've just learned (today) that SQLite respects trailing spaces when comparing two character fields. I.e. 'SQLITE' <> 'SQLITE ' Is this behavior intentional? Neither SQL Server nor Oracle do this. Just curious as to why it works this way.
Because respecting the actual contents of the string is the better way to do things.
The strings 'SQLITE' and 'SQLITE ' are not the same string. Just as the strings 'SQLITE' and 'sqlite' are not the same string. A computer language is more logical, predictable, and easy to use when a test for equality or inequality actually treats every distinct value as distinct. If trailing spaces were supposed to be insignificant for an equality test, then it should not be possible to define a string value containing trailing spaces at all.
Moreover, treating 'SQLITE' and 'SQLITE ' as not being the same string also is consistent with the SQLite philosophy, because it means SQLite has fewer exceptions to be concerned with in a simpler set of rules, and also not having to check lengths and space pad before each compare also makes the code simpler, and less buggy, and it saves CPU cycles. A value equality test is a very common and fundamental thing to do in a DBMS, and bloating that will have a big performance impact.
-- Darren Duncan ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------
