Question on Characetr Encoding to UTF-8

[Joe San]

Guys,

I have an Oracle database which is unfortunately not UTF-8 encoded. When I tried to insert some special characters (for example: the German umlauts) it was not able to recognize and so it replaced those characters with a question mark which is obvious. Now I would like to convert that to UTF-8 encoding. I guess there is a pre defined process of doing that with Oracle. Just wanted to know if there is any other alternative way without changing the Oracle migration approach.

[Joe San]

Any ideas?

[Fatih Keles]

Here Oracle documentation 9i explains character set migration. I am not a dba but changing character set of an existing database seems to be a serious job. I would prefer the export import method if I have to do that on my own. Since character set is strictly related how data is stored, it may result with data corruption and data loss.

Regards,
Fatih.

[Joe San]

I know the docs from Oracle and the different migration approach they propose. But I would like to know of other alternatives which could be done at the application level.

[Joe San]

Looks like the Oracle migration approach would become too tedious. Need to think of alternatives.

[Fatih Keles]

Hi Joe,

As I was looking for something else then I came across this one at AskTom site. As Tom says it is not possible which makes sense.
What is your current NLS_CHARACTERSET? User claims some java guys have done this. Interesting issue? I want to give a try if I find a non-UTF8 database.

Regards,
Fatih.

[Joe San]

Oracle has this WE8ISO8859P1 and the NLS-CHARSET is set to the German locale which means that support is there for the German umlauts. But what about other European characters? So it is for this reason, I wanted to have a simple solution instead of taking the complicated way of migrating the Oracle database to UTF-8

[Joe San]

I was wondering if storing the unicode characters as it is in the database would bring any drawbacks or sideeffects?

For example:, the Turkish characters unicode representation is


Turkish Unicode Characters
ç --> \u00E7
ğ --> \u011F
ı --> \u0131
ş --> \u015F
ö --> \u00F6
ü --> \u00FC
Ç --> \u00C7
Ğ --> \u011E
İ --> \u0130
Ş --> \u015E
Ö --> \u00D6
Ü --> \u00DC 

Would storing this as it is in the database have any side effects? For example. ğozil would be stored as \u00E7ozil in the database.