DSXchange

Hi,
I'm using Character Set as "EBCIDIC" in seq file stage, nls as UTF-8. It is reading the ! in mainframe ebcidic file as ] character. I checked in below site that datastage import/export operator converts ! as ].

https://www.ibm.com/support/knowledgece ... ities.html

Is there any other way of reading this as ! symbol itself in datastage. I got the update from mainframe team as they are using CCSID 037. I tried dd command, and that is also converting to ] symbol. I tried changing NLS setting IBM037, ASCL_MS1252 but not working.

I got lots of jobs. If I need to do this transformer stage then I need to apply this transformation in all the columns and in all the jobs.

Is it really a text file or is it more of a complex flat file with stuff like packed decimal fields in it?

Sounds like your data is not 100% pure text to me.

Seq file stage does do the EBCIDIC to ACII conversion, but not well if it has odd fields like packed decimal and such.

*plink plink, 2 cent drop*

There are different EBCDIC code pages. Your mainframe folks have clued you in by telling you they are using CCSID 037. The ! character is an example that would translate differently if using, e.g., CCSID 500.

You need to experiment with the EBCDIC NLS code pages and use whatever matches up to CCSID 037.

Without specifying a code page, you're getting a default EDCDIC to ASCII conversion, which apparently is not what you want.

Mike

IBM037 would have been my first guess too, but since it apparently didn't work out for you, I'd go with ibm-37-s390 (EBCDIC United States) as my second guess.

Mike

Its a plain text file. No packed decimal in it. I put a server job , and read as binary data type and used ASCII function in the transformer stage and it is working fine. I can see the ! symbol in the output file. But don't know how to get this in parallel job.

I'm getting ! characater in output now . I used "Unicode" in column property and used IBM037. Thanks all :D