SPUM4 Vocabulary

[Former Member]

Dear All,

We are in the process of sap upgrade 4.6 C à ECC6.0 Combine Upgrade and Unicode Conversion (CU and UC).

The active languages in our system UK,CS,DE,EN,FR,HU,IT,DLA,PL,NL,NO,PT,SK, RU,ES,FI,SV.

As per sap note 756535(SAP Vocabulary for Vocabulary Maintenance for SPUMG & SPUM4), we have auto assigned Vocabulary delivered by SAP.

And then as per sap note: 871541(Frequently used text patterns), we have imported SAP delivered frequently used text patterns extracted from 11 European and Asian languages.

Also after assigning all vocabulary provided by sap, there are still 65000+ words which are not covered by the any of the provided language vocabulary.

Is there any sap standard way/notes/guides to handle this situation and assign language for remaining 65000 words as this information is not provided in the guide.

Thanks & Regards

Rahul

[nils_buerckel]

Hi Rahul,

There are various methods to maintain the vocabulary with the correct language keys (which are also listed in the CU&UC guide):

a. Import of existing vocabulary (see SAP note 756535)

b. Language dependent scan (optional - would skip it for PRD)

c. Creation and use of hints (see SAP Note 1034188)

d. Statistical methods based on technical attributes (not available for 4.6c)

e. Statistical methods (see SAP note 756534 / not available for 4.6c)

f. Use of vocabulary patterns (see SAP note 871541)

g. Manual Maintenance (made by native speakers)

For all automatic or semi-automatic methods it is important to check the result with native speakers. This is especially valid for methods b) u2013 f). The manual maintenance should be executed by native speakers.

The sequence of the methods listed above can be used as a hint how to do sequencing during the actual maintenance (of course this is just a proposal - variations are possible). It is pretty clear, that e.g. the manual maintenance should be minimized and therefore should be executed at the end of the list. On the other hand side, existing vocabulary (SAP vocabulary and / or customer vocabulary from earlier scans) should be added at the beginning of the vocabulary maintenance.

For some of the methods described in b. to f. it is possible that the results are not as expected and words are assigned wrongly. Therefore customers need to check the outcome after each step and it is advisable to decide whether to skip the step or make use of it.

In your case it might be useful to have a detailed look into point c: Hint creation

Vocabulary words caused by Company Code dependent tables can be usually assigned to a specific language via company code. Therefore in case of many vocabulary words from this type of tables, it might be useful to create according hints. SAP Note 1034188 lists some more fields (next to company code BUKRS), which are capable to be used for assigning the language via hints. However it needs to be verified with application experts that the selected attributes can be used properly to assign a code page.

Usually it makes sense to create hints only for tables, which contribute a high number of words in the vocabulary. This can be identified via report UMG_VOCABULARY_STATISTIC or UM4_VOCABULARY_STATISTIC

It is recommended to execute UMG_DISTRIBUTE_ASSIGNED_LANGS (see SAP Note 938374) after using the hints.

Best regards,

Nils Buerckel

SAP AG

[sudheer_ellath2]

Hi All,

This thread is very informative, but while executing those steps we are  facing some problem. We are doing MDMP conversion and end up with around 15,000 words and as part of reducing it, we tried to perform those steps mentioned in 756534 where the program UMG_UPDOWN_CHARSTAT  is going for shot dump as well as for 1034188, there is no clear indication where we will get those pogroms to be executed for each table.

It will be really helpful if anyone who applied these notes provide some lights in to it.

Thanks in advance

[nils_buerckel]

Hi,

regarding the short dump for the program it is difficult to say anything without knowing details. However I recommend to concentrate rather on the hints than looking at the automatic assignment.

SAP note 1034188 serves as info for the hint assignment. However this is NOT an automatic process.

Here you need to do the following steps:

- Identify tables which contribute a lot of words to the vocabulary (e.g. via UMG_VOCABULARY_STATISTIC)

- Check whether these tables have fields as indicated in SAP note 1034188 (fields with which the code page can be ambiguously identified)

- As a result there should be a list of tables where it is feasible to create hints

- Hints have to be created individually per system based on the points outlined above

The creation of hints is described in the Unicode conversion guide as attached to SAP notes 551344 or 928729.

Best regards,

Nils Buerckel