Proper Data type for transferring XML from ABAP to XI

Question

Dear Colleagues,
I have a question on how to wrap XML data, that itself contains an encoding
declaration, so that it will be properly transferred to XI.
XML Generation: "always use X"
I have searched a lot on SDN and other resources but could not find
information, describing how a web service should wrap unicode-data when
making it available to XI. The XML processing group emphasizes that all generation of XML should be done using X-bases (HEX) datatypes in ABAP and preferably using UTF-8 as ancoding. This by itself is not a problem.
But transferring UTF-8-XML - through which interface?
When a webservice provides an XML Stream and wants to transfer it to XI then there is a bit of a paradox situation, because it has to choose between using an X-based or Character-based type for its interface
-> as Binary data?
- when using X-based data types in the interface, I have control of the character encoding, but I expect that the XML-information will probably be wrapped in base64 encoding and then the XI will receive a base64 encoded
XML-block instead of "plain" XML.
I did not find any reference information on how the soap interface of XI and the WAS behave when transferring X-based data, however.
-> as Character data?
- on the other hand when using Character-based data types (e.g. String), I have no control of the codepage being used. I have then only the choice to generate the default encoding of the application server. Maybe this encoding
will be transformed later to a different encoding, but usually during this process the header, which declares the encoding in use will not be adapted as well.
So if I have a unicode machine, it will generate UTF-16 - based XML. I have no other choice, if I use a character-based type for output. And this will also be stated in the header of the XML stream. If the XML stream
leaves the system it will probably be transformed to UTF-8, but the XML-header will still state the content to be UTF-16, because the header will probably not be auto-converted together while adapting the Character
encoding.
Because of these issues I am quite insecure, on how to generate and wrap UTF-8 - based XML, so that I can properly hand it over to XI, ensuring that
it stays UTF-8 and that XI can really unwrap it automatically.
Does anyone have a tip in which type to wrap XML-based data on the ABAP stack so it will be usable in XI?
Thank you very much for your attention and time.
Yours
Christian Steinert

Answer 1

Something like this should work (with a little tweek I imagine)
This would be in the 'Command on Target':
OdiFtpGet "-HOST=<%=odiRef.getInfo("SRC_DSERV_NAME")%>" "-USER= '<%=odiRef.getInfo("SRC_USER_NAME")%>'" "-PASSWORD= <%=odiRef.getInfo("SRC_PASS")%>" "-REMOTE_DIR=<%=odiRef.getInfo("SRC_SCHEMA")%>" "-REMOTE_FILE=<%=odiRef.getTable("L", "COLL_NAME", "W")%>" "-PASSIVE_MODE=YES"
Set the Technology of the 'Command on Source' as File, and the technology of the 'Command on Target' as ODI Tools
Edited by: Bos on Jun 30, 2011 11:53 AM