Re: extracting double byte characters

Hi - I am working in a Japanese environment and have implemented the situation you describe below. In a database configured to store double byte characters - Japanese Kanji characters are represented in 2 bytes - whereas the 26 character alphabet and other characters that can normally be represented in ASCII are stored in one byte. Oracle uses - what they call - a shift-in/shift-out control character to represent the double byte character - for example, here is how a data field might look:

<shift-out control char> < Kanji ... kanji ... kanji> <shift-in control char> <a, b, c, d, ....>

If you use SQL*NET to communicate between dbases configured for single and double byte - Oracle is smart enough to know that the NLS set up is different between instances. Characters which come in to the single-byte instance that are double byte will be automatically represented using some default graphic (I heard you can change the graphic - but I'm not sure how. My system prints a box when when it encounters this). Single-byte to Single-byte (ie - normal ascii) are represented as they are in printable ASCII.

Note also - if your data field is all ascii chars in the double byte instance, there will be no shift-in/shift-out control chars. Those control chars are only created when a double byte char needs to be represented.

• Troy

Steven WM Wong wrote:

> If we implement a double byte database to capture transactional information
> in Chinese as well as some information in ASCII (English + numeric data),
> then is it possible for us to extract English (ASCII) information out from
> the double byte database and upload it into a single byte database (another
> instance)? How are the opposite, i.e. can data from a single byte database
> be extracted into a double byte database? And work to enable this? What are
> additional considerations?
>
> Thanks a lot.
> Stevn Wong
Received on Sun Jan 23 2000 - 18:54:42 CST