Loading a UTF-8 reports into z/OS Ondemand

[alexwlchan]

Recently, we have a UTF-8 report that we would like to upload to z/OS Ondemand because we would like to support reports in multi languages (Traditional Chinese and English). I have tried 2 approaches

1. Loading reports using generic indexer
- I can load the reports into Ondemand using codepage 1208 and the UTF-8 text could be display properly
- I have the CC defined in first column (i.e. "1" (HEX value 31 for UTF-8)) even I enabled the CC = YES and CCTYPE = ANSI but this still could not split pages. Would there be any way to split the pages?
- It also shows error when I enable "Large Object" in application with ARS1176E. I think that should be "page" related.

2. Loading reports using OS-390 indexer
- I tried many ways but still failed with ARS5471E and ARS5419E
- Any idea what setting that we can do for the reports to be loaded with OS-390 indexer?

[Justin Derrick]

Can you post examples of the error messages (to save everyone from having to look them up themselves)?  Help us to help you. 

-JD.

[Ed_Arnold]

Loading a NODX document via the OS/390 Indexer is a special case.

The developer who wrote the OS/390 Indexer had an article on NODX in the 1Q2012 newsletter under the z/OS section. 
                                                                         
That article is still 100% relevant to NODX indexing and nothing in it has changed.                                                             
                                                                         
  http://www-01.ibm.com/support/docview.wss?uid=swg27023970              
   
Ed                                                                     

[Ed_Arnold]

Just a couple notes:

1) The generic index format has no supported way of specifying the page-level indexes.

2) One characteristic of the ACIF and OS/390 indexers is that the coordinates to find stuff on the page are in bytes, however UTF-8 is a MBCS.  (Multi-Byte Character Set)  It can be difficult to specify coordinates when every character before the ones you are interested in can be 1-4 bytes in length.  What may work for one load may not work for the next.

Ed