[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: [no subject]
Thanks!
Regards,
Alex Yiu
------_=_NextPart_001_01C475C5.A89411E6
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Diso-8859-1">
<TITLE></TITLE>
<META content=3D"MSHTML 6.00.2800.1400" name=3DGENERATOR></HEAD>
<BODY text=3D#000000 bgColor=3D#ffffff>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>Hello,</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>(this=20
mail is about assign's native data handling in the for-each case and =
about $var=20
syntax as a possible helper)</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>The=20
discussion on this issues is a bit calmed down, however I run across one =
major=20
point, and want to share my findings (especially with those of you =
thinking=20
about using JXPath for assign queries):</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>To=20
actually dynamically create the result of a parallel/sequential =
for-each, one=20
could use the XPath createPathAndSetValue() method with an predicate =
for a=20
new Element:</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>Sample=20
instance of a message variable:</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2><root></FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2> =20
<level1><level2>AA</level2></level1></FONT></SPAN=
></DIV>
<DIV><SPAN class=3D248405622-29072004>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2> =20
<level1><level2>BB</level2></level1></FONT></SPAN=
></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></root></FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2><assign></FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2> =20
<from>bb</from></FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2> =20
<to variable=3D"msgVar" part=3D"root" =
query=3D""/root/level1[2]/level2"=20
/></FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></assign></FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>Can be=20
implemented with the follwoing JXpath method:</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>ctx.setValue("/root/level1[2]/level2", =
"bb")</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>The=20
assign could also be used to actually create a new =
element:</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>ctx.createPathAndSetValue("/root/level1[3]/level2",=20
"CC")</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>However, this legal XPath construct has a few requirements for =
the=20
predicate (the [3] on the axis):</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>a)=20
must be a number (a function returning a string does not=20
work)</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>b) the=20
expression must be a context indeendend location =
path</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>The=20
given sample works, however, if we want to use a BPEL =
counter/instance-number,=20
this will look like:</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>/root/level1[bpws:getVariableData("counter",=20
"count")]/level2</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>This=20
does not work with current JXPath implementations, since extension =
functions are=20
not context independend (I guess this can be changed) AND it requires =
the=20
function to return a numbr instance.</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>I=20
wonder if the BPEL spec should therefore require engines =
to</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>-=20
provide a jxpath parser which allows getVariableData() in create=20
path</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>-=20
provide a xpath extension function which is type =
aware</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>BTW:=20
using the $counter variable will solve the issue for JXPath, it asumes =
variables=20
to be context independend:</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>/root/level1[$counter]/level2</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2>If you=20
are curious, here is the result of a small JXPath test (using my own=20
DOMAbstractFactory and some sample extension =
functions):</FONT></SPAN></DIV>
<DIV><SPAN class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff =
size=3D2></FONT></SPAN> </DIV>
<DIV><SPAN class=3D248405622-29072004>
<P><FONT face=3DArial color=3D#0000ff=20
size=3D2>dom.getVal(/root/level1/level2)=3DAA<BR>dom.getVal(/root/level1[=
2]/level2)=3DBB<BR>dom.getVal(/root/level1[last()]/level2)=3DBB<BR>dom.ge=
tVal(count(/root/level1))=3D2.0<BR>dom.getVal(/root/level1[count(/root/le=
vel1)]/level2)=3DBB<BR>dom.getVal(/root/level1[count(/root/level1)-1]/lev=
el2)=3DAA<BR>dom.getVal(concat('hallo',count(/root/level1)))=3Dhallo2<BR>=
dom.getVal('/root/level1/level2')=3D/root/level1/level2<BR>dom.getVal(eck=
i:str1()+ecki:str2())=3D3.0<BR>dom.getVal(concat(ecki:str1(),ecki:str2())=
)=3D12</FONT></P>
<P><FONT><FONT face=3DArial><FONT color=3D#0000ff><FONT size=3D2><SPAN=20
class=3D248405622-29072004>OK, enough basic XPath syntax, now the real=20
fun:<BR></SPAN></FONT></FONT></FONT></FONT><FONT><FONT =
face=3DArial><FONT=20
color=3D#0000ff><FONT=20
size=3D2>dom.getVal(/root/level1[ecki:str1()]/level2)=3DAA<BR>dom.getVal(=
/root/level1[ecki:str2()]/level2)=3DAA</FONT></FONT></FONT></FONT><FONT=20
color=3D#0000ff><SPAN class=3D248405622-29072004><FONT face=3DArial=20
size=3D2> &nbs=
p; =20
<- return type is string!<BR></FONT></SPAN><FONT face=3DArial=20
size=3D2>dom.getVal(/root/level1[number(ecki:str2())]/level2)=3DBB</FONT>=
</FONT><FONT=20
color=3D#0000ff><SPAN class=3D248405622-29072004><FONT face=3DArial=20
size=3D2> <- casted to right =
type<BR></FONT></SPAN><FONT=20
size=3D2><FONT =
face=3DArial>dom.getVal(/root/level1[ecki:int2()]/level2)=3DBB<SPAN=20
class=3D248405622-29072004> &nbs=
p; <-=20
now correct return type</SPAN></FONT></FONT></FONT></P>
<P><FONT face=3DArial color=3D#0000ff size=3D2><SPAN =
class=3D248405622-29072004>Now we=20
try to modify an existing node:<BR></SPAN></FONT><FONT face=3DArial =
color=3D#0000ff=20
size=3D2>ctx.createPathAndSetValue(/root/level1/level2, XX) =
done.<BR></FONT><FONT=20
face=3DArial color=3D#0000ff =
size=3D2>dom.getVal(/root/level1[1]/level2)=3DXX<SPAN=20
class=3D248405622-29072004> &nbs=
p;  =
; =20
<- yes it worked</SPAN><BR></FONT><FONT face=3DArial><FONT =
color=3D#0000ff=20
size=3D2>dom.getVal(/root/level1[2]/level2)=3DBB<SPAN=20
class=3D248405622-29072004> &nbs=
p;  =
; =20
<- affected only the first node</SPAN><BR></FONT></FONT></P>
<P><FONT><FONT><SPAN class=3D248405622-29072004><FONT face=3DArial =
color=3D#0000ff=20
size=3D2>Now we want to create a new one (first check to see that =
there is=20
no 3rd level1 element:)<BR></FONT></SPAN></FONT></FONT><FONT><FONT =
face=3DArial=20
color=3D#0000ff=20
size=3D2>dom.getVal(/root/level1[3]/level2)=3Dorg.apache.commons.jxpath.J=
XPathException:=20
No value for xpath: /root/level1[3]/level2<BR></FONT></FONT><FONT=20
face=3DArial><FONT color=3D#0000ff><FONT=20
size=3D2>ctx.createPathAndSetValue(/root/level1[3]/level2, YY)=20
done.<BR>dom.getVal(/root/level1[3]/level2)=3DYY<SPAN=20
class=3D248405622-29072004> &nbs=
p;  =
; =
=20
<- now a third node exists</SPAN></FONT></FONT></FONT></P>
<P><SPAN class=3D248405622-29072004><FONT face=3DArial size=3D2>Now we =
try the same=20
with a number expression;<BR></FONT></SPAN><FONT color=3D#0000ff><FONT =
face=3DArial=20
size=3D2>ctx.createPathAndSetValue(/root/level1[ecki:int2()+ecki:int2()]/=
level2,=20
ZZ) failed: org.apache.commons.jxpath.JXPathException</FONT><SPAN=20
class=3D248405622-29072004><FONT face=3DArial=20
size=3D2>: <BR> =20
</FONT></SPAN></FONT><FONT face=3DArial color=3D#0000ff =
size=3D2>Exception trying to=20
create xpath /root/level1[ecki:int2()+ecki:int2()]/level2;<BR><SPAN=20
class=3D248405622-29072004> =20
</SPAN>JXPath can only create a path if it uses exclusively the child:: =
and=20
attribute:: axes and has no context-dependent =
predicates<BR></FONT></SPAN><SPAN=20
class=3D248405622-29072004><FONT face=3DArial color=3D#0000ff=20
size=3D2></FONT></SPAN></P></DIV>
<DIV></SPAN></SPAN><FONT face=3DArial size=3D2>Mit freundlichen =
Gr=FC=DFen<BR>Bernd=20
Eckenfels<BR></FONT><FONT face=3DArial size=3D1>Chief =
Architect<BR>--<BR>SEEBURGER=20
AG - Edisonstr.1 , D-75015 Bretten, Germany<BR>Fax: +49 (0)7252 96-2400 =
- Phone:=20
+49 (0)7252 96-1256<BR><A=20
href=3D"mailto:b.eckenfels@seeburger.de";>mailto:b.eckenfels@seeburger.de<=
/A> - <A=20
href=3D"http://www.seeburger.de/"; =
target=3D_blank>http://www.seeburger.de</A></FONT>=20
</DIV></DIV></DIV>
<BLOCKQUOTE>
<DIV class=3DOutlookMessageHeader dir=3Dltr align=3Dleft><FONT =
face=3DTahoma=20
size=3D2>-----Original Message-----<BR><B>From:</B> Alex Yiu=20
[mailto:alex.yiu@oracle.com]<BR><B>Sent:</B> Tuesday, May 11, 2004 =
7:33=20
PM<BR><B>To:</B> ygoland@bea.com<BR><B>Cc:</B> Ron Ten-Hove; Danny van =
der=20
Rijn; wsbpeltc; Alex Yiu<BR><B>Subject:</B> Re: [wsbpel] Issue 103 - =
draft=20
proposal<BR><BR></FONT></DIV><BR>Yaron, <BR><BR>Thank you for reading =
the=20
proposal in details. <BR>I totally agree with you that it will still =
take a=20
number of emails to get all details of this proposal flushed out =
right. :-)=20
<BR><BR>Summary: <BR>---------------------------------<BR>Regarding to =
QNAME=20
collision, that issue has been at the back of my mind for a while. I =
have been=20
putting it in a lower priority in my thinking process. I also want to =
solve=20
the collision issue. <BR><BR>I agree with mostly with your point =
"<I>1) Enable=20
BPEL variables to be defined using both simpleTypes and complexTypes =
(I have=20
been meaning to propose this anyway).</I> <I>Then require that XML =
schemas=20
auto-generated from WSDL messages be anonymous ...</I>" <BR><BR>For =
your point=20
"<I>2) Introduce the bpel:partName attribute ...</I>" ... I am not =
sure that=20
is necessary.<BR>---------------------------------<BR><BR><BR>Detailed =
reply=20
to your feedback (which is very constructive):<BR><BR>(I)<BR>"<I>1) =
Enable=20
BPEL variables to be defined using both simpleTypes and =
<BR>complexTypes (I=20
have been meaning to propose this anyway).</I>" <BR><BR>If my analysis =
(below)=20
is right, we can solve the QNAME collision problem without formal =
introducing=20
ComplexType into variable declaration. We may no need to introduce =
complexType=20
to solve this issue. Let's handle complexType issue separately. (I am =
still=20
open to that idea in general.)<BR><BR>(II)<BR>Also under "1)", =
"<I>Then=20
require that XML schemas auto-generated from WSDL messages be =
anonymous=20
...</I>"<BR><BR>I think I like this idea. We can avoid QNAME collision =
by=20
generating anonymous QNAME for XML Schema Element corresponding to a =
WSDL=20
Message Type. <BR><BR>Given that adopting this mechanism, I would =
remove item=20
(4) in my proposal ("messageType" =3D> "messageElement"). Keeping =
messageType=20
attribute would be fine. <BR><BR><variable name=3D"fooMsg"=20
messageType=3D"mywsdl:myFooMsgType" /> <BR><BR>By using messageType =
attributes, the BPEL translator knows it will generate an anonymous =
QNAME of=20
an XML Schema Element for "mywsdl:myFooMsgType". =
<BR><BR>(III)<BR>"<I>2)=20
Introduce the bpel:partName attribute ...</I>"<BR><BR>I guess there is =
no=20
collision for part name in general. Because, for a message part, we =
are=20
generating a local element declaration and definition. The part name =
does not=20
constitute a QNAME. The part name is a non-qualified local element=20
declaration. Using your example, "Details" and "MoreDetails" are not =
qualified=20
elements. (Just like "product" element can have a "name" =
subelement=20
which is a string of maxLength =3D 50, while "customer" element can =
have a=20
"name" subelement which is a string of maxLength =3D 40.)<BR><BR>(IV) =
XPath=20
Semantic Clarification:<BR><BR>If you look back to my proposal in item =
(1):=20
<BR>"<I>The value of a BPEL variable, which appears as =
Variable-Reference in=20
XPath 1.0, corresponds to a node-set of exactly one XML node in XPath =
1.0=20
semantics. The XML node corresponds to the <SPAN=20
style=3D"FONT-WEIGHT: bold">element node (not document node)</SPAN> of =
related=20
BPEL variable ...</I>"<BR><BR>There is a big difference between =
element node=20
or document node in terms of how XPath look like. <BR><BR>If we go for =
a=20
document node, your XPath will be right:=20
<BR>"$BV/a:MT/a:Part/a:Element"<BR><BR>If we go for an element node, =
the XPath=20
in the proposal will be right:<BR>"$BV/a:Part/a:Element"<BR><BR>Here =
is a=20
little experiment that I run on XMLSpy to illustrate=20
=
this:<BR><BR>emp_test.xslt<BR>----------------------------------<BR><?=
xml=20
version=3D"1.0" encoding=3D"UTF-8"?><BR><xsl:stylesheet =
version=3D"1.0"=20
xmlns:xsl=3D<A class=3Dmoz-txt-link-rfc2396E=20
=
href=3D"http://www.w3.org/1999/XSL/Transform";>"http://www.w3.org/1999/XSL=
/Transform"</A>><BR> =20
<xsl:output method=3D"xml" version=3D"1.0" encoding=3D"UTF-8"=20
indent=3D"yes"/><BR> <xsl:template=20
match=3D"/employeeTag"><BR> =20
<xsl:variable name=3D"empVar" select=3D"." =
/><BR> =20
<test><BR> =
=20
<test1><xsl:value-of =
select=3D"$empVar/@id"=20
/></test1><BR> =20
<test2><xsl:value-of =
select=3D"$empVar/fname"=20
/></test2><BR> =20
<test3><xsl:value-of =
select=3D"name($empVar)"=20
/></test3><BR> =20
</test><BR> =20
=
</xsl:template><BR></xsl:stylesheet><BR>---------------------=
-------------<BR><BR>emp.xml<BR>----------------------------------<BR><=
;?xml=20
version=3D"1.0" encoding=3D"UTF-8"?><BR><employeeTag=20
id=3D"101"><BR> =20
<fname>John</fname><BR> =20
=
<lname>Dole</lname><BR></employeeTag><BR>--------------=
--------------------<BR><BR>Execution=20
Result:<BR>----------------------------------<BR><?xml =
version=3D"1.0"=20
encoding=3D"UTF-8"?><BR><test><BR> =20
<test1>101</test1><BR> =20
<test2>John</test2><BR> =20
=
<test3>employeeTag</test3><BR></test><BR>--------------=
--------------------<BR><BR>If=20
my proposal got accepted, we need to change Section 14.3 accordingly a =
little=20
bit in the spec. <BR><BR>(V)<BR>From reading your appendix, I am glad =
that you=20
also think getting rid of part names is a good idea.<BR>We are more or =
less on=20
the same page =
there.<BR><BR><BR>Thanks!<BR><BR><BR><BR>Regards,<BR>Alex=20
Yiu<BR></BLOCKQUOTE></BODY></HTML>
------_=_NextPart_001_01C475C5.A89411E6--
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]