On Thu, Jun 01, 2000 at 01:00:16AM -0800, guy ruth hammond wrote:
> Andreas Jung wrote:
>
> > This is what I mean "space wasting". There is no need store *every* character of
> > a string to be stored in 2 bytes. This is a really problem. Why should
> > I need 20GB of disks to store just 10 GB of data that contains
> > very few 2 byte characters like german umlauts....
>
> Andreas,
>
> If this is a problem, you'll have to take it up with the International
> Standards Organization, not with Oracle. I'm sure Oracle's engineers
> have the skill to create a 9-bit character set, but they have no control
> over what UTF8 is. :0)
>
> If you are really worried, use HTML-style entity encoding and 8-bit
> characters. Here is an example of what I mean (it can be cut out and
> pasted into an HTML file and viewed in a web browser).
>
> <HTML>
> Soufflé
> </HTML>
>
> I couldn't think of a word with an umlaut in it offhand.
That's not the point...when I store UTF8 text in a varchar2
fiels each character allocates 1 byte for characters with code <127 and
2 bytes for the others. In CLOB Oracle 8i stores *EVERY* character
in fixed 2 bytes format (UCS-2). I believe in the skills of the
Oracle programmers but I don't see why where is a difference between
in storing UTF8 in varchar2 and clob fields :-) It's a question
of space and a question of much money for disks and raid arrays.
Received on Thu Jun 01 2000 - 04:35:27 CDT
