On Wed, May 31, 2000 at 02:44:06AM -0800, guy ruth hammond wrote:
> Andreas Jung wrote:
> >
> > Dear all,
> >
> > I am running a Oracle 8i database with UTF8 character set. I store text files
> > inside a CLOB (extra lob segment). I can see that Oracle always needs two times
> > the space to store the text inside the LOB segement compared to the size of the
> > originial file. Is this a general problem or is this problem related to the
> > UTF8 character set ?! It seems to me that Oracle need 2 bytes for every
> > character although most of the text is pure Ascii.
>
>
> UTF8 is a 16-bit character set. ASCII is an 8-bit character set,
> which lacks the NLS capabilities of UTF8. So storing in UTF8 will
> be twice as large as storing in ASCII - and if you used plain
> characters, the additional capability of UTF8 is wasted.
I am aware of this...but UTF8 is *NOT* a fixed character set.
Only characters with codes > 127 are stored in 2 bytes. Oracle
does not use 2 bytes when you store UTF8 in varchar2 fields.
So this is space wasting.
Received on Wed May 31 2000 - 07:53:10 CDT
