The results are correct. The expression `loc_town REGEXP '[A-ZŁa-zł]'` will return *true* for any value of `loc_town` which contains the specified character in any position within it. The expression `loc_town REGEXP '([A-ZŁ][a-zł])\w+'` will return *true* for any value of `loc_town` which contains the specified characters followed by one or more words. The appropriate REGEXP pattern to match (return *true*) for strings composed exclusively of [A-ZŁa-zł] characters and spaces (and is at least 1 character long) would be '^[A-ZŁa-zł ]+$'. ^ means *start of field* and $ means *end of field*