The TRANSLATE function translates a character value expression
    from one character set to another compatible character set.

    The characters in the char-value-expr are translated, character-
    by-character, to the character set indicated by the translation
    name. If a direct translation exists for a character, it is
    replaced by the equivalent character in the translation character
    set. If there is no direct translation for a character, it is
    replaced by the space character of the translation character set,
    as shown in the example using the TRANSLATE function.

    For example, the Kanji character set contains traditional
    Kanji characters, Katakana characters, ASCII characters,
    and Roman characters that are ASCII characters encoded in 2
    octets. If a Kanji char-value-expr is translated using the
    RDB$KATAKANA translation name, those 2-octet Kanji characters
    that have an equivalent in the 1-octet Katakana character set
    are translated. The other characters are replaced by the Katakana
    space character.

    If a character in the source character string is not compatible
    with the target character set, SQL substitutes a space character
    for that character.