-----BEGIN PGP SIGNED MESSAGE-----

> And good idea of using subtypes for VARCHAR and CHAR fields that
> can indicate short binary data. This enables new data type aliases:
> VARBINARY(n) and BINARY(n) (someone asked about octet(4) for ip
> adresses?).
> Unfortunately, this can introduce incompatibility when software
> that assumes every VAR/CHAR is textual tries to convert it to text.

CHAR(4) CHARACTER SET OCTETS

Works perfectly well for storing/retrieving and moving around (in
local variables in triggers and stored procedures) things like IP
addresses in network format.

I also routinely use CHAR(8) CHARACTER SET OCTETS for temporary
storage (local variables) of RDB$DB_KEYs.

Is there really a need to invent a BINARY ?
Is it part of any SQL standard ?

Are there reasons why the above should not work ?

- --

Olivier Mascia, om@..., Senior Software Engineer
T.I.P. Group S.A., www.tipgroup.com, Director




-----BEGIN PGP SIGNATURE-----
Version: PGP Personal Privacy 6.5.8

iQEVAwUBOgq4f6IodNUVZIMJAQHUqgf+NEXGaC7PpSiMcI7g+WpgozexQgVCODC2
XHT3KCOtzgdW2wMH/zxc3bVz+8VDTs07Yb4OA0H4My/8wehzpCJN/VhhY9vi/2sI
Tf6qYkp5IWjUAaNkbuMowkYmPaNc0dQAU75VVWHoC6928i3W4DqGszq00bUbzlii
yhO4/LyOG9flm2njATqgEZKXVvfhV5z63Zr1He4HZCTs6uxZ3JGcxvZxiWWe1Fyf
XJ3UUbDTu8pqov5WxrfMfhOcLA3+JF6y9ZErLtH4ADHJ8IltIoaR5YYAcVWVoG7G
0IC5YFX9vzLl3y75PsOYacwqfmw/u2x2OVRLVQnuI99VLtsuX5ukBA==
=5W6Z
-----END PGP SIGNATURE-----