Hello Mark,
Thank you for your deteiled reply which solved my confusion completely.
Yes the first choice you pointed out is proper to initialize a complex
variable. I had falsely understood the relation between "literal
variant of copy" and "TII -> EII".
As to my example, the first copy should be:
<copy>
<from><literal>
<svc01:elemB xmlns:svc01="http://example.com/service01.wsdl";>
<svc01:elemC/>
</svc01:elemB>
</literal></from>
<to>$_b</to>
</copy>
rather than
<copy>
<from><literal><svc01:elemC
xmlns:svc01="http://example.com/service01.wsdl"/></literal></from>
<to>$_b</to>
</copy>
Hope it helpful to others.
Thanks again.
On Mon, Apr 6, 2009 at 9:49 PM, Ford, Mark <mark.f...@ll.mit.edu> wrote:
> The BPEL spec only addresses copy operations. As such, the to-spec for the
> copy operation must result in a target node that exists. A TII -> EII copy
> operation will never result in the creation of a child element in the target
> node. The result will be the target element having a single child text node,
> not an element node.
>
> What you have run into is a valid exception being thrown because your second
> copy operation is targeting a node that does not exist in the variable $_b.
> After the first copy operation, the contents of $_b are:
>
> <svc01:elemC/>
>
> You are attempting to copy $_c into a child of $_b that doesn't exist. For
> example:
>
> <svc01:elemC>
> <svc01:elemC/> <!-- This is the node you're targeting -->
> <scv01:elemC>
>
> The quote from the spec that you referenced pertains to the literal variant
> of the copy operation. With this variant, the from-spec must produce a single
> EII or a single TII. In your case, you are producing a single EII which is
> valid. Therefore it is not a TII -> EII copy operation, it's actually an EII
> -> EII copy operation.
>
> Your choices are (in no particular order):
>
> * initialize the entire target variable using the complete element in the
> literal from-spec
> * use extensions to dynamically create missing target nodes
> * use Xquery expressions in the from-spec to construct the complex
> variable dynamically in a simple EII -> EII copy op to update the whole
> variable or parts of it.
>
> On 4/6/09 4:19 AM, "ZHAO Wenfeng" <zhaowenf...@gmail.com> wrote:
>
> According to WS-BPEL(Version 2.0) specification, it seems that the "TII ->
> EII" type of copy can be used to initialize a complex variable - at least to
> create its first children. The section "8.4.2. Replacement Logic of Copy
> Operations" says(P.70):
> "To replace the destination content:
> If the destination is an EII, all [children] properties (if any) are
> removed and the source content TII is added as the child of the EII. "
>
> But ODE 2.0 seems not to comply with it because in my case the assignment:
> <copy>
> <from><literal><svc01:elemC
> xmlns:svc01="http://example.com/service01.wsdl"/></literal></from>
> <to>$_b</to>
> </copy>
> <copy xmlns:svc01="http://example.com/service01.wsdl";>
> <from>$_c</from>
> <to>$_b/svc01:elemC</to>
> </copy>
> will incur selectionFailure:
> ERROR - GeronimoLog.error(104) | Assignment Fault:
> {http://docs.oasis-open.org/w
> sbpel/2.0/process/executable}selectionFailure,lineNo=98,faultExplanation={http:/
> /docs.oasis-open.org/wsbpel/2.0/process/executable}selectionFailure: No
> results
> for expression: {OXPath10Expression $_b/svc01:elemC}
>
> But alternatively, using the XPath extention fuction insert-as-last-into()
> provided by ODE, the above initialization can be accomplished.
>
>
> The version of ODE I use is ODE 2.0 Build #87 (2009-1-8 2:25:04). The
> declaration of the variables is as:
> <variables xmlns:ws0="http://example.com/service01.wsdl";>
> ... ...
> <variable name="_a" element="ws0:elemA" />
> <variable name="_b" element="ws0:elemB" />
> <variable name="_c" element="ws0:elemC" />
> <variable name="_a2" element="ws0:elemA2" />
> </variables>
> And the schema is:
> <schema targetNamespace="http://example.com/service01.wsdl";
> xmlns="http://www.w3.org/2001/XMLSchema";
> elementFormDefault="qualified">
> ... ...
> <element name="ResultInfo">
> <complexType>
> <sequence>
> <element name="elemA">
> <complexType>
> <sequence>
> <element name="elemB">
> <complexType>
> <sequence>
> <element name="elemC" type="string"
> maxOccurs="unbounded"/>
> </sequence>
> </complexType>
> </element>
> </sequence>
> </complexType>
> </element>
> <element name="elemA2" type="string"/>
> </sequence>
> </complexType>
> </element>
> </schema>
> The complete file set is in the attachment.
>
>
> If this disagreement in ODE is not a bug but a intentional design, I guess
> the reason is that, according to WS-BPEL, even if it is supported, only the
> first child can be created and the creation of other children must appeal to
> some extension mechanism. See "8.4 Assignment"(P.62):
> "The fifth from-spec variant returns values as if it were a from-spec that
> selects the children of the <literal> element in the WS-BPEL source code.
> [SA00038] The return value MUST be a *single* EII or Text Information Item
> (TII) only."
>
> Am I right?
>
>
> Thanks & Regards
> Wenfeng
>
>
> ----------
> ZHAO Wenfeng
> http://www.bupt.edu.cn
>
>
>
>
> --
> Mark Ford
> MIT Lincoln Laboratory
> 244 Wood Street
> Lexington MA 02420
> (781) 981-1843
>
Regards
Wenfeng
---------------------
Zhao, Wenfeng
http://www.bupt.edu.cn
