Oracle Database Discussions

Please note
1. this is a forum and not Oracle Support: you cannot ask for "urgent" answer.
2. character set migration has nothing to do with ASM: you should post to right forum 3052
3. CSSCAN does not convert data: it reports data that does not need conversion and data that needs specific conversion.

Hi,
Thanks for your sugesstion and reply.

CSSCAN is a read only utility - it does not make any changes

http://docs.oracle.com/cd/E11882_01/server.112/e10729/ch12scanner.htm#i1005939

The CSALTER code will perform the actual conversion, assuming all of the prereqs are met - the time time taken will depend on the data present in the database that needs to be converted. The time can only be estimated by performing a test convert on test hardware that is similar or identical to production

HTH
Srini

Thanks a lot srini for your reply.

One more doubt here. I would like to create new database with required character set and import old data into database. Will it show correly in new database.

It depends - on what the current characterset is and how data is stored in the current database - pl provide details on what issue you are trying to solve

HTH
Srini

Hi Srini,

Please find the current set of db here. Existing DB is supporting Western Languages. Our requirement is need to store chinese character in DB.


PARAMETER VALUE
NLS_LANGUAGE AMERICAN
NLS_TERRITORY UNITED KINGDOM
NLS_CHARACTERSET WE8ISO8859P15
NLS_DATE_LANGUAGE AMERICAN
NLS_NCHAR_CHARACTERSET AL16UTF16
NLS_LENGTH_SEMANTICS BYTE
NLS_NCHAR_CONV_EXCP FALSE
NLS_RDBMS_VERSION 11.1.0.7.0
As per my analysis, I found two solutions for this problem.
1) by running CSSCAN utility need to covert DB but as I reffered there is a chance of loosing or truncating data.
2) Create new DB with required set up and importing existing DB data.

Kindly suggest best solution to solve this problem.

Thanks
Hareesh Reddy.

Please note that CSALTER will NOT convert user data. It will only convert CLOBs in the data dictionary (i.e. system columns). It will also change the character set declaration but it will not re-encode user data for the new character set. Only the DMU (http://www.oracle.com/technetwork/products/globalization/dmu/overview/index.html) and export/import can convert the database. The DMU does it in place, the export/import requires data to be moved out from the database and reinserted (or moved to another database).

Also, unless the database needs to support Oracle 8.x clients, do not migrate to the deprecated UTF8. Migrate to AL32UTF8.

-- Sergiusz

Hi Wolicki,
Thanks for your reply and advice. Hope DMU tool is availabel in 11.2 version but we are using 11.1. Do I need to upgrade to 11.2 otherwise any other alternative is there.

Thanks
Hareesh Reddy

Please [url http://docs.oracle.com/cd/E39404_01/doc/index.htm]read the fine manual.
[url http://docs.oracle.com/cd/E39404_01/doc/doc.12/e36716/ch2migrasteps.htm#BABDHEAC]DMU Overview of Requirements and Security Considerations [url http://docs.oracle.com/cd/E39404_01/doc/doc.12/e39375/toc.htm]DMU Release notes

You can find all DMU supported database releases and platforms here: http://www.oracle.com/technetwork/products/globalization/dmu/learnmore/configurations-333837.html

Only 11.2.0.3 is supported on all platforms (except EBCDIC mainframes). Selected older releases are supported on selected platforms provided an appropriate patch is installed.


-- Sergiusz

Hi ,
Currently we are using Unix AIX operating system which is having 64 bit configuration. Kindly let me how can I safely change character set without losing any data. Can you guide me which will support Chinese character in application. Please find current database setup details in the thread it self.

Thanks
Hareesh Reddy

Edited by: 992356 on Mar 11, 2013 10:47 AM

You should use AL32UTF8 - http://docs.oracle.com/cd/E11882_01/server.112/e25494/create001.htm#ADMIN13343

HTH
Srini

Agreed. Always use AL32UTF8 unless you have a very good reason to use one of the legacy character sets.


-- Sergiusz

Hi Srini/all
Thanks a lot for your guidence and help to resolve character set issue. I have run csalter utility and noticed that few of the application tables are impacted. Kindly let me know how can I handle Truncation, Lossy data.

Pl see these MOS Docs

Csscan Output Explained [ID 444701.1]
Solving Convertible or Lossy data in Data Dictionary objects reported by Csscan when changing the NLS_CHARACTERSET [ID 258904.1]

HTH
Srini

Did you check the scan reports for what tables are impacted and what the lossy data are?

It is you who knows the application data best and should be able to find out how you can hande conversion exceptions.

Earlier in this thread you got several suggestions on the DMU tool. How did that turn out?


Read the doc for a description of e.g. Individual Exceptions and the two types of app data exceptions.
http://docs.oracle.com/cd/E11882_01/server.112/e10729/ch12scanner.htm#i1007101