Skip to Content
avatar image
Former Member

SLT from DB2 to BW with CCSID on DB2

We have a requirement to extract data from a DB2 database (backend to a JD Edwards ERP) into SAP BW using SLT. We were able to successfully connect to the database and replicate the datasources. However, the proposed datatypes for the datasource were matching the source datatypes and were labeled as 'RAW'. We performed a sample extraction into BW, which was successful. However, upon looking at the data - all the char/txt fields were extracted as hexadecimal values.

We took a closer look at the DB2 tables, and found out that the fields are CCSID enabled (mostly 65535). I am told that this is fairly standard for DB2 databases.

The question is - how is this issue dealt with in cases of SLT/BW? Is there a way to decode the values in the source before SLT replicates this information? If not, is we do bring in the raw encoded data, is there a way within SLT or BW/HANA to decode this information as it loads into BW?

Here are the source system details:

ERP Application: JD Edwards World A9.4

Database System & Version: IBM DB2 V7R2

Operating System Version: OS400 V7R2

Server Hardware: IBM Power 7

Add comment
10|10000 characters needed characters exceeded

  • Get RSS Feed

2 Answers

  • May 15 at 08:59 AM

    Columns of type CHAR or GRAPHIC with a CCSID of 65535 are representing binary or raw data, i.e. no character conversion takes place when accessing this data. Valid character data would be tagged with an EBCDIC CCSID (such as 37 or 500) or a Unicode CCSID like 13488. Did you check with JDE/Oracle why columns that contain text data are marked with a CCSID that represents binary data?

    Kind regards,

    Christian Bartels.

    Add comment
    10|10000 characters needed characters exceeded

  • avatar image
    Former Member
    May 14 at 07:22 PM

    Can anyone from the SAP SLT team shed some light?

    Add comment
    10|10000 characters needed characters exceeded