Message Mapping : Play around with target structure
Perhaps there may be many ways to manage all the above requirements but sometimes it becomes like using a cannon to kill a fly! So here we go with our favorite message mapping and see how to play around with tag name, occurrence, namespace and prefix at target side. Also we will try to pass value to structure node, seems weird? Let’s see how it is possible with simple user defined function.
In first step to access target field in udf, we take help of getParameter method of Container with parameter STRUCTURE_NODE that returns access to target field/node to which the UDF is being directly mapped to. The node needs to be type casted to structure/leaf as per the type of target field.
- StructureNode node = ((StructureNode) container.getParameter(“STRUCTURE_NODE”));
Applicable to fields having child or sub-fields.
- LeafStructureNode node = ((LeafStructureNode)container.getParameter(“STRUCTURE_NODE”));
Applicable to fields having no child or sub-fields.
Next step involves using node methods as per the requirements.
- String getQName()
Return the name of target field.
- void setQName(String newName)
Set the target field name with input parameter newName.
- void setNSDeclarations(String decl)
Set the namespace declaration and attribute with input parameter decl.
e.g decl = ” attribute1=\”a1\” attribute2=\”a2\” xmlns:prefix=\”namespace1\””
- void setOccurances(int min, int max)
Set the occurrence of target field with input parameter min and max.
- void setPreValue(String prevalue)
Add the string prevalue unescaped just after start tag of target field. Applicable only for structure field.
- void setPostValue(String postvalue)
Add the string postvalue unescaped just before end tag of target field. Applicable only for structure field.
For example, lets say we have the source and target structure as shown below:
Field Name | Occurence | Field name | Occurence |
---|---|---|---|
Source | 1..1 | Target | 1..1 |
Rows | 0..Unbounded | Items | 0..Unbounded |
Key | 1..1 | Element | 1..1 |
Value | 1..1 |
And the requirements are..
- The Element field name should get changed with value in Key field and should be filled with Value.
- Output xml should have only 2 items assuming that it is coming in a particular order from source.
- The namespace and its prefix should be Modified to “http://company.com” and pfx respectively.
- Assign value “list of top 2 company” to Target node.
Source instance and expected result xml are as shown below.
Source Instance | Result |
---|---|
<?xml version=“1.0” encoding=“UTF-8”?> <ns0:Source xmlns:ns0=“http://test.com“> <Rows> <Key>CompanyA</Key> <Value>abc corp</Value> </Rows> <Rows> <Key>CompanyB</Key> <Value>xyz info</Value> </Rows> … <Rows> <Key>CompanyZ</Key> <Value>123 ent</Value> </Rows> </ns0:Source> |
<?xml version=“1.0” encoding=“UTF-8”?> <pfx:Target xmlns:pfx=“http://company.com“> list of top 2 companies <Items> <CompanyA>abc corp</CompanyA> </Items> <Items> <CompanyB>xyz info</CompanyB> </Items> </pfx:Target>
|
And here comes the mapping part..
- /ns0:Target=setNSDeclarations(const(value=pfx), const(value=”http://company.com“)
UDF setNSDeclarations would fix namespace to http://company.com , fix prefix to pfx and assign value “list of top 2 companies” to Target node.
public String setNSDeclarations(String prefix, String nmspace ,Container container) {
StructureNode node = ((StructureNode) container.getParameter(“STRUCTURE_NODE”));
node.setNSDeclarations(” xmlns:” + prefix + “=” + nmspace); // notice the space before xmlns:
node.setQName(prefix + “:Target”);
node.setPreValue(“list of top 2 companies”);
return“”;
}
- /ns0:Target/Items=setOccurences(/ns0:Source/Rows, const(value=2))
UDF setOccurences simply set the max occurence of Items to 2 to ensure top 2 Rows are mapped to Items.
public String setOccurences(String sourcenode, int maxOccur, Container container) {
StructureNode node = ((StructureNode) container.getParameter(“STRUCTURE_NODE”));
node.setOccurences(0, maxOccur);
return “”;
}
- /ns0:Target/Items/Element=setTargetFieldName(/ns0:Source/Rows/Key, /ns0:Source/Rows/Value)
UDF setTargetFieldName dynamically changes the tag name Element.
public String setTargetFieldName(String targetFieldName, String targetFieldvalue, Container container) {
LeafStructureNode node =((LeafStructureNode) container.getParameter(“STRUCTURE_NODE”));
node.setQName(targetFieldName);
return targetFieldvalue;
}
At end, please note that the code discussed above uses undocumented classes/methods that are subject to change in future without notice. In short, use at your own risk as any related issues will not be supported by SAP
Reference :
Nice..info..How did you found the API for these ? Are there published some where ? if so can you add that link in above, so that it will be helpful ..
Regards
Rajesh
Thanks Rajesh. I came across those undocumented methods while writing UDF in developer studio with reference to mapping api ( com.sap.xi.mapping.tool.lib_api.jar, com.sap.xpi.ib.mapping.lib.jar) that we use for java mapping. Auto completion with ctrl space is the way forward for any further experiment 🙂
Regards,
Sunil Chandra
Good one Sunil. I use to develop Java mapping for such scenarios, now I can achieve the same using UDFs. Keep blogging..
Thanks
Rakesh Sharma
I agree with Rakesh, thank you for this blog! 🙂
Hi Sunil,
Thanks for sharing, it's always good to explore some undocumented features 🙂
Regards,
Greg
Hi Sunil,
This is very helpful blog. Expect more from you.
Regards
Anupam
Hello Sunil,
This is top class. Would be very helpful for consultants who prefer graphical mapping.
cheers,
Ambrish
Nice one Sunil
Hi Sunil,
Thanks for a wonderful blog. please let me know can we create Target Structure at run
time ?
for exp :- I want to add extra record or element in target structure at run time ..
Thanks & Regards
Rudra
Nice Blog. How are you passing integer to UDF ? Is this supported in Integration Builder PI 7.30?
Thank you for that blog - I`m going to test it right now 🙂
Hi Sunil,
Nice blog, as you said Any PI developer favorite mapping would be Message Mapping and even mine 😎
Hi Sunil.
Very good blog.Really helpful.Thanks for sharing this information.
Regards,
Abhi
Hi Sunil ,
So nice, Useful document. This doc helped me in my recent projects..................
Thanks for sharing.
Regards,
Vijay Kumar K.V.N
Hi Sunil... Thanks for the blog. Really informative. Can you help me with my requirement ?
I need to create additional fields in the target(apart from those present in MT) dynamically in mapping.
How would I do that?
Find the below screenshots for reference.
Very helpful doc for me. Thank you for sharing it.
BR,
Rashmi
i'm getting one requirement like it is a grouped interface we have contracts,payments,division order,tax and regulatory combine all these interfaces need to implement in one interface.
Scenario is REST--->PO---->IDOC
For all interfaces we have one common IDOC structure available on Target side and this interface need to run hourly basis daily.
How can i implement this interface.I need your suggestions ?
BR,
venkat
Thanks for this solution as it helped me a lot to avoid cumbersome xslt mapping which did not work either due to conversion from < > " to < > and " values causing the end result to throw error reference not in prolog, THis resolved my solution in a jiffy.
THANKS ONCE AGAIN for SOME GREAT WORK PRODUCED HERE FOR OTHERS TO RESOLVE THEIR OWN PROBLEMS
Danie Luttig
Hi Sunil,
This blog is very helpful.
Keep on posting blogs like this!!