Re: How to fix invalid DBCS data?

Paul wrote:
One of the features offered by the IBM Midrange platform
is an ability to build and use custom character translation tables
as an additional supplement to the base support
offered by a particular CCSID.

You might want to check with the Japanese subsidiary
to see if they followed this path and if so whether they could
e-mail the relevant table or tables to you.
There is an import method you need to follow correctly to have the tables
update your system to make the DBCS data reference the proper characters.


Paul -

The DBCS fields in the original Japanese files use CCSID 930.

Only certain fields in certain records have problems - for example, the
description field in records 1, 5, and 100 in a given file may have a
problem, but all of the other records are Ok.

It is more like they have missing shift-in and shift-out characters - if the
user brings up the specific records in question in the original database and
updates them with their original green-screen program which uses a display
file with CCSID 930 fields and saves the value, then the field converts
properly once we re-import the data into the new application, and it is
viewable through the new application program.

I'm gathering more information for a subsequent post to this thread, with
more details about the process that we are following to do the conversion of
the data.

Regards,
Steve