Problems with UNICODE-characters from Oracle-database

ray.wurlod · Post by **ray.wurlod** » Fri Nov 25, 2011 1:41 pm

Don't trust what View Data shows you in regard to "un-American" characters. Use a viewer that can display the Euro symbol, accented characters, etc. properly.

paultechm · Post by **paultechm** » Fri Nov 25, 2011 4:51 pm

Apply stringtoUstring(columnvalue,'UTF-8') and make column datatype as nvarchar

View data wont give the correct result ,please check the final out put (file or table)

Thanks
Paul Joseph

coface · Post by **coface** » Mon Nov 28, 2011 3:22 am

paultechm wrote:Apply stringtoUstring(columnvalue,'UTF-8') and make column datatype as nvarchar

Hi Paul,
Our DataStage-developer will try this suggestion. I will come back with the result...

Rgds
Jan

coface · Post by **coface** » Mon Nov 28, 2011 5:10 am

Hi Paul,
sorry to say: but that didn't help

paultechm · Post by **paultechm** » Mon Nov 28, 2011 5:43 am

whats the value looks like in the table after applying this funcation .

Make sure that datatype is in nvarchar

coface · Post by **coface** » Mon Nov 28, 2011 6:48 am

the value looks like an upsidedown questionmark...
Our datastage-developer just told me that this problem occurs only with parallel jobs - but not with server-jobs. Since I'm not a datastage developer I don't have any idea why this makes such a difference. Maybe it's a bug in parallel jobs?
I have opened a SR at IBM - hopefully they are able to tell us why this error occurs only with parallel jobs...

Rgds
Jan

paultechm · Post by **paultechm** » Mon Nov 28, 2011 7:40 am

Whats your source and datatype used for reading? Are you checking in view data or table?

coface · Post by **coface** » Mon Nov 28, 2011 8:01 am

the source is a database table (as described it's an external table which directly reads from a flatfile).
We have tried all kinds of datatypes: varchar (with and without extended unicode, nvarchar). None of those worked...

paultechm · Post by **paultechm** » Mon Nov 28, 2011 8:11 am

Use DRS satge to read the data as varchar in transformer convert using the function and change datatype as nvarchar .

chulett · Post by **chulett** » Mon Nov 28, 2011 8:36 am

People, please...

Stop the love affair with the Reply with quote option. Sure, quote someone when the need arises to make a point or ensure we all know who / what is being responded to but to quote everything all the time?

Use the big orange Reply to topic 'button'.

coface · Post by **coface** » Tue Nov 29, 2011 3:04 am

... as a matter of fact this problem only occured with parallel jobs but not with server jobs. Therefore the developer decided to use a server job as a workaround.
Thanks to all who discussed this issue with me..