OnDemand User Group

Support Forums => Report Indexing => Topic started by: DDP021 on September 18, 2020, 06:20:09 AM

Title: Character translation
Post by: DDP021 on September 18, 2020, 06:20:09 AM

Hello,

We have an application sending a mainframe file to CMOD.   One of the characters is a | (x'4f' when turning on hex in the mainframe dataset).  When it loads to CMOD, its showing as a ! (exclamation point).  Does anyone know why CMOD is doing this?

Title: Re: Character translation
Post by: Justin Derrick on September 18, 2020, 06:57:13 AM

Can you give us a run-down of the whole process?  Where the file is produced, how it's transferred, your Application definition (codepages, user exits), is it transformed on the way out of CMOD, which client are you using, is there any translation on the client side, etc, etc...

Codepages are like timezone libraries...  We do our very best to avoid scratching the finely polished exterior to see the horrors just below the surface!

-JD. 

Title: Re: Character translation
Post by: DDP021 on September 18, 2020, 07:28:15 AM

Hi Justin,

We are running CMOD 10.1.0.5.  Running Content Navigator. Version: 2.0.3 Build: icn203.813.478. 

This file is created as a mainframe dataset and then loaded into CMOD via ACIF indexer.  The dataset itself is allocated as FBA with a record length of 261.

The CMOD definition is defined with a Codepage of 500.  CC=Yes, RECFM=Fixed, LRECL=261, CC Type=Ansi

Both Client and Navigator are showing the same character.  Looking at it, this is the only character not showing correctly.  Not aware of any type of translation being used.

Title: Re: Character translation
Post by: Darrell Bryant on September 21, 2020, 06:57:50 AM

It appears to me that the dataset is encoded in CCSID 37 but is being loaded as CCSID 500. Code point X'4F' is different between the two with the difference you describe.

https://www.ibm.com/support/knowledgecenter/SSEPEK_10.0.0/char/src/tpc/db2z_codeptdiffebcdic.html

Title: Re: Character translation
Post by: DDP021 on September 21, 2020, 07:19:24 AM

Changing the CodePage to 37 worked...Now showing | and not !...How did you decipher what Codepage the mainframe dataset was using?  Appreciate the help!!!!

Take care

Dave

Title: Re: Character translation
Post by: Darrell Bryant on September 21, 2020, 07:27:50 AM

Hi Dave,

Based on your description, with only one character being incorrectly displayed, it sounded like an encoding problem to me. When I googled the phrase  code page 500  the first two hits described differences from other EBCDIC code pages, including the problem you were experiencing.  The second hit being more readable, I included it in my reply.

Best Regards,
Darrell

Title: Re: Character translation
Post by: DDP021 on September 21, 2020, 07:31:23 AM

Gotcha!...I hadn't looked at your link you included yet....I know Justin and mentioned CodePage maybe being a issue!  I appreciate the help!!!!!

Title: Re: Character translation
Post by: Justin Derrick on September 21, 2020, 07:53:34 AM

Heh...  You two were up early this morning.  :)

I'm glad to see it was as simple as changing the codepage...  +1 to Darrell!

-JD.