data to Unix-Oracle as garbage

[fahad]

Hi all,
i want to move arabic data from text file to Unix-Oracle database. oracle character set is AR8ISO8859P6 (to have arabic data) but there is no NLS set like that in Datastage.
i tried with many NLS setting on the Oracle stage of the job but either give error or deliver data as garbage (question marks).

does any one face a problem like that before? where can i search for NLS new patches from Datastage if there is any.

thanks in advance

[roy]

Hi,
you could start by reading the NLS manuals.
they also cover the ones available.
in any case you can always build one yourself.

IHTH,

[sachinkc]

Just giving some notes from experience here rather than answering your question directly:
I'm sure you must have checked it before, but is the data is the text file of similar character set?
Also, turning to UTF-8 might be a good idea, but of course, that depends on what direction your company wants to take.
- Sachin

Hi all,
i want to move arabic data from text file to Unix-Oracle database. oracle character set is AR8ISO8859P6 (to have arabic data) but there is no NLS set like that in Datastage.
i tried with many NLS setting on the Oracle stage of the job but either give error or deliver data as garbage (question marks).

does any one face a problem like that before? where can i search for NLS new patches from Datastage if there is any.

thanks in advance

[ray.wurlod]

Arabic is one of the more difficult character sets, particular for sorting and comparing. I don't currently have access to an NLS-enabled DataStage, so can't see what maps are supplied with the product.

What you will need to know (to build your own map) is how your Arabic characters are encoded and how to map from these encodings to Unicode code points.

The process is described in the NLS manual. It may be worthwhile engaging Ascential consulting services to help you to build the map.