Grouper Users - Open Discussion List

[Dominique Petitpierre <>]

Hello,

On 06/17/2013 03:02 PM, Gagné Sébastien wrote:
> The source of the problem is that the grouper DB field is 255 BYTES, not
> 255 CHARS, accent characters use 2 Bytes,  hence the larger value (some
> other like the apostrophe take 3)

This would be true if one uses a database encoding like UTF-8. In
Nicolas' case, the database for Grouper is Oracle DB, with a database
encoding ISO8859-1 (WE8ISO8859P1 in Oracle speak), which stores one
byte for every character.

We can see it with the following SQL request for a display_extension
of exactly one character: "é" (e-acute).

SELECT dump(display_extension,1016)
FROM grouper_groups
WHERE display_extension='é';

Typ=1 Len=1 CharacterSet=WE8ISO8859P1: e9

This result shows that the length of the stored value is 1 byte and the
character is encoded with the byte e9 (hexadecimal) which is the code
of e-acute in the ISO8859-1 character set (cf. "man iso_8859-1" on
Linux).
Reference:
http://www.oracle.com/technetwork/products/globalization/nls-lang-099431.html


So the problem that Nicolas was reporting is not caused by the
database encoding.  (i.e. display_extension has a data_type
VARCHAR2(255 BYTE) and will accept up to 255 iso_8859-1 encoded
characters).
The problem is "up stream", probably in the grouper-ws application
that he was using.

Best regards,
Dominique
--
Mr Dominique Petitpierre, user=Dominique.Petitpierre domain=unige.ch
IT Division, University of Geneva, Switzerland