Problem publishing database contents from non-unicode to unicode system

[Former Member]

Hello everyone!

We just set up a new SAP WAS based on Netweaver 2004 as a unicode system. Out problem now is that we have a content management system on our non-unicode system and that we are publishing the contents via rfc to the WAS unicode system to display the contents online. the contents are stored in our own database tables.

The problem thereby is that many texts pasted from microsoft word contain special characters like bullets, long minus or low-9 quotation marks which are not correctly displayed in the unicode system / on the website. we already found out that it has something to do with the codepage. the sap notes say we should use 1160 instead of 1100 and that the transaction SPUMG would be helpful. but we are not able to select any tables there.

so now we do not know what to do exactly. do we have to change something in our non-unicode system or do we have to conversion in our unicode system. and what happends if content containing special microsoft word characters is published after the spumg conversion? do we have to to this frequently?

We would be glad if anyone could help.

Thanks a lot!

[martin_fuchs2]

Hi Stefan,

I hope I understand your issue...

You are trying to publish contents from a non-Unicode-system to a Unicode system (website).

The character you mention are not part of the [ISO standard codepages for Latin-1|http://msdn.microsoft.com/en-gb/goglobal/cc305167.aspx]. That confuses me, because these characters cannot be stored in non-UC-systems directly.

How is that data being stored in the non-UC system?

How is the system communication set up and what are the RFC settings?

Compact answer to your questions: Unicode conversions are done once per non-UC system and transaction SPUMG is being used during preparation of this process.

Best regards,

Martin Fuchs

SAP AG

[Former Member]

Hi Martin,

thanks for your quick answer.

You got me right. We have a local non-Unicode SAP HCM Netweaver 2004 system running a self-developed web based content management system / wiki. The texts entered in the bsp application are stored in a string field in our database table. Actually we publish the contents to a WAS 6.20 non-Unicode system with the same database tables to provide the content via BSP for the public. Everything is working fine including the special characters.

Now we want to replace the WAS 6.20 non-Unicode system by a new WAS 7.0/2004 Unicode system. But when publishing the contents via the same RFC function module to the new system the special characters seem to be damaged. We are not able to replace them with abap commands and when they are displayed on the website we only see "boxes".

If I get you right we have to run SPUMG on our nw 2004 non-unicode productive hcm system, right? but isn't there a danger to damage existing contents?

Best regards,

Stefan

[martin_fuchs2]

Hi Stefan,

first of all: SPUMG is only to be used in the process of a full Unicode conversion. The conversion guides can be found here (links on right side): [Unicode Overview page in SDN|https://www.sdn.sap.com/irj/sdn/i18n?rid=/webcontent/uuid/c09cb763-f841-2a10-bcb1-d4cc8973e1eb] [original link is broken];

I guess the problem arises from your RFC destination which is now sending Unicode by default. Please check whether switching to non-Unicode in the RFC configuration changes the received data.

Best regards,

Martin

[Former Member]

Hi Martin,

I think that was the the right hint for the problem's solution. The current settings for the target system in the tab "MDMP & Unicode" for the RFC-Connection in SM59 are: non-Unicode and MDMP inactive. Therefore the special characters are replaced by a #.

Unfortunately I'm not able to change the settings because they are deactivated. What can I do to change this?

Thanks again for your help.

[martin_fuchs2]

Did you switch to CHANGE mode? (pen button)

[Former Member]

Yes of course, i can edit all the other settings there is just this one tab inactive.

[nils_buerckel]

Hi Stefan,

in a Non-Unicode system the "Unicode" switch in SM59 is always deactivated - the reason for that is that the Non-Unicode system does not do a conversion - this is always executed on the Unicode side (in case of ABAP - ABAP interfaces).

The reason for your problem is in deed that a Non-Unicode Western European abap system supports 1100 only (ISO-8859-1). But the MS code page 1252 (SAP notation 1160) has additional characters (please have a look at note 552612 for details). These additional characters are not supported on the 1100-system. Therefore the conversion is like 1100 --> Unicode - all additional characters cannot be converted properly.

Unfortunately the only solution to this problem I see is a Unicode conversion of your Non-Unicode system ...

Best regards,

Nils Bürckel

SAP AG