when to run SPUMG for production unicode conversion

[Former Member]

We will be executing the actual MDMP Unicode conversion for the live Production system. We intend to run the SPUMG scan and vocab/reprocess log language assignment one week before the export.

As the production system is still operational, there will new new words created after the completion of the SPUMG. How are these new words handled, will the new unassigned words be in the SUMG lists ?

[peter_dzurov]

I think you will need to do "Update worklist" and then run at least scans for "INITIAL" tables in SPUMG right before conversion, it is a part of additional preparation steps in non-unicode system. Please check and proceed according to guide. All unassigned words should appear in SUMG after conversion then.

[Former Member]

I think the "update worklist" function only scans for new tables created by support packages, upgrade PREPARE phase and new custom table. It will not update the vocabulary and insert new words, am i correct ?

[nils_buerckel]

Hi Thom,

R3load is "looking" at all language independent db table entries and detects words with special characters within these entries.

Now assuming there is a new entry, then there are two possibilities:

a) all words (with special characters) in this entry already appeared in other old entries (rows) of this table and they are assigned in the vocab

b) There are completely new words (for this table), which do not exist in the vocab

In case a), R3Load is able to detect the source code page and converts the entry correctly.

For case b), R3Load normally uses the fallback code page and these entries will come up in SUMG in the manual repair (assuming that reprocess scans were also finished when the new entry was created).

Regards,

Nils Buerckel

Globalization Services

SAP AG

[Former Member]

hi nils,

if that is the case, then the reprocess scan and index repair can only be done the very last minute ?

after that, freeze on data entry of foreign words ?

[nils_buerckel]

Hi Thom,

the assumption is that there are not too many entries with new words - and these can be handled in SUMG after the conversion via manual repair. A "data freeze" is usually not possible ...

Best regards,

Nils