Re: Doubt UTF8 database

Praveen wrote:

> Hi All,
>
> I have Oracle 8.1.6 on win2000 env. I have already one database with
> AMERICAN AMERICA.WE8ISO8859P1 charset. And i created one more database
> with UTF8 charset. If i see in my registry, the variables have
> following values..
>
> HKLM\Software\Oracle\NLS LANG=NA
> HKLM\Software\Oracle\ALL HOMES\ID0\NLS LANG=NA
> HKLM\Software\Oracle\HOME0\NLS LANG ERICAN AMERICA.WE8ISO8859P1.
>
> If client app(which is going to be developed in java) try to access
> this database will it work? or if there is anything is missing in my
> registry.
>
> Please help me,
>
> Praveen

The oracle clients don't care what charset the database server has as long as the client charset is a full subset of the server's. Server we8iso8859p1, client us-ascii => works fine, but not reversed.

Then a session is created, a bunch of statements like alter session set nls charset=...., number delimiters and ore. That means that the connecting process on the server adapts the settings from the client, if possible.

Server UTF8, client could have any set of we8iso8859p1-17, traditional Chinese, north-korean, russian, greek ....

The edge is, that your english or us client possibly isn't able to display the north-korean text, and will be shown as reversed ????

Worse is, that many believe that an 'international' windows client might

be able to select Chinese charset - but if it is not a Chinese version windows - it wont work.

Remember that with utf8 charset on the server, some characters like ? ?and others (i cant type due to charset settings) can expand up to 3 bytes. All of a sudden a varchar2(300) holds only 100 characters, if they are all 3 byte types. Some charset are fixed 3 byte and some are variable 1 to 3 bytes. Check the doc, nls lang support.

/svend Received on Sat Jul 12 2003 - 18:26:17 CEST