problem with migration to AL32UTF8 Options

[makaruku]

Hi,
I have a problem with character set migration to AL32UTF8 in Oracle database. My database version is 11.2.0.3 and I’m using Database Migration Utility to Unicode. The most problematic thing is that I have two different chararacter sets in database. NLS_CHARSET is EE8ISO8859P2 and system data is in this charset, the second one is EE8MSWIN1250 and all application data is in that one because it was imported without conversion to ISO. I set in DMU that current charset is ISO, assumed is MSWIN and target is AL32UTF8. I managed to resolve all conflicts except for one – data in table SYS.SOURCE$ need conversion. I don’t know what to do with this one. Could someone help me with that? Is there another way to migrate database to UTF? Oracle support told me that using DMU tool is the only supported method.

[burleson]

Hi Makaruku,

>> Is there another way to migrate database to UTF?

Well, one sure-fire way is to export the data with vanilla SQL.

CODE

spool mytab.dat
select col1||','||col2 from mytab;
spool off;

At that point, it's just a plain flat file.

At that point it can be imported into a new table with a different NLS using SQL*Loader. . . .

[makaruku]

Hi Burleson,
we can't migrate entire database this way because it is too large (almost 1TB). DMU has problem only with SOURCE$ table. I have got one idea to solve our problem but don't know if it works. I was planning to export data from SOURCE$ table (to sql file or as You say with vanilla SQL), truncate it, make migration to UTF using DMU and after migration import exported data to SOURCE$ table in new charset. Is it good idea?

Hi Makaruku,

>> Is there another way to migrate database to UTF?

Well, one sure-fire way is to export the data with vanilla SQL.

CODE

spool mytab.dat
select col1||','||col2 from mytab;
spool off;

At that point, it's just a plain flat file.

At that point it can be imported into a new table with a different NLS using SQL*Loader. . . .

[burleson]

Hi Makaruku,


>> I was planning to export data from SOURCE$ table (to sql file or as You say with vanilla SQL), truncate it, make migration to UTF using DMU and after migration import exported data to SOURCE$ table in new charset. Is it good idea?

Yeah, that's how I would do it!

Check with Oracle tech support, see what they say:

httyp://support.oracle.com

Good Luck!