SPUMG with CHARACTERS?

[Former Member]

Hi Gurus

After running the SPUMG we get thousands of 2 character Vocabulary entries..

For example:

þu2014

©*

¹£

»u2013u2013

Which seem to be just a random combination of NON ALPHA NUMERIC CHARACTERS?

We have 3 codepages to choose from for the language assignment.

My question is what code page should I assign if I only see NON ALPHA NUMERIC CHARACTERS?

Kind Regards

Lawrence Brown

[nils_buerckel · Accepted solution]

Hi Lawrence,

looks like characters from a Non-Latin code page like Asian, Russian or Greek ...

If you have such a code page, please check with native speakers (logging on with the according language) if these are valid words.

If you do not have such code pages or the byte combination does not make sense in all your three code pages (when logging on with the correct languages), then these could be words from code pages you do not have in the system.

One typical case where this occurs is in case of ECC 6.0, if you do not use Russian. Please have a look at note 1275317.

If you are not able to assign the words, then I would recommend to leave the words empty in the vocabulary and check in the reprocess scan, which table entries caused the problem.

However it is always possible to repair the entries with SUMG after the Unicode conversion.

Best regards,

Nils Buerckel

SAP AG

[Former Member · Accepted solution]

Hello Nils,

Thanks alot for answering. We are using the Russian codepage so we will proceed as you suggested with a native speaker.

Kind Regards

Lawrence

[Former Member]

Hi,

It's depend up on your system codepage

So,You can check the code page in the following way then assign the codepage:

Table : ARCH_CP

Function Module : SCP_GET_CODEPAGE_PROPERTIES " check if the codepage of the file is a unicode codepage

SCP_GET_CODEPAGE_NUMBER " no unicode system -> Check if MNLS support (MDMP)

check in the standard program : LARCHF02

You can check ur system codepage in Function Module: SYSTEM_FORMAT, SYSTEM_CODEPAGE

Regards,

Sathish

[Former Member]

HI Santish,

This is the process to convert to UNICODE, and this is a MDMP system so there are 3 code pages.

Thanks again

Lawrence

[Former Member]

Put them in exception list and continue to finish all the scans first .