Is it possible to preprocess xml source within the same XSLT Stylesheet? - xslt

Within the same XSLT (2.0) Stylesheet and transformation I would like to:
1) first preprocess the whole XML Datasource (Add a attribute
with a specific calculation to certain elements)
and then
2: transform the changed XML Datasource with the sylesheet's templates.
How can I achieve this? A code example would be nice?

Yes it is possible. One possibility would be to to proceed as follows:
<xsl:template match="/">
<!-- store the modified content in a variable -->
<xsl:variable name="preprocessed.doc">
<xsl:apply-templates mode="preprocess" />
</xsl:variable>
<!-- process the modified contents -->
<xsl:apply-templates select="$preprocessed.doc/*" />
</xsl:template>
<!-- first pass: sample process to add an attribute named "q" -->
<xsl:template match="*" mode="preprocess">
<xsl:copy>
<xsl:copy-of select="#*"/>
<xsl:attribute name="q"><xsl:number count="*" level="any" /></xsl:attribute>
<xsl:apply-templates mode="preprocess" />
</xsl:copy>
</xsl:template>
<!-- "normal" processing of the modified content. It is able to used the newly processed attribute. -->
<xsl:template match="*[#q <= 5]">
<xsl:copy>
<xsl:apply-templates />
</xsl:copy>
</xsl:template>
<xsl:template match="*"/>
In the first pass, an attribute is added, counting the element in the input XML.
In the processing, we only retain the elements having the value of the q attribute set a number less or equals to 5.

Related

XSLT: create node if does not exists

I know similar questions are already there but none of them seem to work for me.
So shortly, I have XML file with tag "Lokal" that in most cases does not appear but it should. Not making things easier: I also need to change a name of "Lokal" to let's say "Lokal_test". My goal is modify node name(if exists) or create it and rename (if does not exists).
Data from XML will be imported to MS Access data so they need to match perfectly with table...
Sample XML:
<Dane>
<InformacjeOWpisie>
<DaneAdresowe>
<AdresGlownegoMiejscaWykonywaniaDzialalnosci>
<Budynek>3a</Budynek>
<Wojewodztwo>podlaskie</Wojewodztwo>
</AdresGlownegoMiejscaWykonywaniaDzialalnosci>
</DaneAdresowe>
</InformacjeOWpisie>
<InformacjeOWpisie>
<DaneAdresowe>
<AdresGlownegoMiejscaWykonywaniaDzialalnosci>
<Budynek>8r</Budynek>
<Lokal>2</Lokal>
<Wojewodztwo>mazowieckie</Wojewodztwo>
</AdresGlownegoMiejscaWykonywaniaDzialalnosci>
</DaneAdresowe>
</InformacjeOWpisie>
</Dane>
Desired output:
<Dane>
<InformacjeOWpisie>
<DaneAdresowe>
<AdresGlownegoMiejscaWykonywaniaDzialalnosci>
<Budynek>3a</Budynek>
<Lokal_test/>
<Wojewodztwo>podlaskie</Wojewodztwo>
</AdresGlownegoMiejscaWykonywaniaDzialalnosci>
</DaneAdresowe>
</InformacjeOWpisie>
<InformacjeOWpisie>
<DaneAdresowe>
<AdresGlownegoMiejscaWykonywaniaDzialalnosci>
<Budynek>8r</Budynek>
<Lokal_test>2</Lokal_test>
<Wojewodztwo>mazowieckie</Wojewodztwo>
</AdresGlownegoMiejscaWykonywaniaDzialalnosci>
</DaneAdresowe>
</InformacjeOWpisie>
</Dane>
This question(XSLT: create node if not exists seemed to be the awnser to my problems but when trying to use it does not work.
Not sure why?
<xsl:template match="InformacjeOWpisie/DaneAdresowe/AdresGlownegoMiejscaWykonywaniaDzialalnosci/Lokal">
<Lokal_test>
<xsl:apply-templates select="#*|node()" />
</Lokal_test>
</xsl:template>
EDIT:
When I get rid of parent Lokal_test dissapears. I use below code to say "bye bye" to parent:
<xsl:template match="InformacjeOWpisie/DaneAdresowe/AdresGlownegoMiejscaWykonywaniaDzialalnosci">
<xsl:apply-templates select="#*|node()" />
</xsl:template>
<xsl:template match="#*|node()">
<xsl:copy>
<xsl:apply-templates select="#*|node()"/>
</xsl:copy>
</xsl:template>
<xsl:template match="AdresGlownegoMiejscaWykonywaniaDzialalnosci/Budynek">
<xsl:copy>
<xsl:apply-templates/>
</xsl:copy>
<xsl:choose>
<xsl:when test="exists(following-sibling::Lokal)">
<Lokal_test>
<xsl:value-of select="following-sibling::Lokal"/>
</Lokal_test>
</xsl:when>
<xsl:when test="not(following-sibling::Lokal)">
<xsl:element name="Lokal_test"/>
</xsl:when>
</xsl:choose>
</xsl:template>
<xsl:template match="Lokal"/>
You approach was right, but incomplete. You only created the new Local_test element.
So try these two templates in combination with the indentity template:
<!-- Handles the replacement of the 'Lokal' element -->
<xsl:template match="AdresGlownegoMiejscaWykonywaniaDzialalnosci/Lokal">
<Lokal_test>
<xsl:apply-templates select="node()|#*" />
</Lokal_test>
</xsl:template>
<!-- Creates a new 'Lokal_test' element if no 'Lokal' element exists -->
<xsl:template match="AdresGlownegoMiejscaWykonywaniaDzialalnosci[not(Lokal)]">
<xsl:copy>
<xsl:apply-templates select="node()/following-sibling::Wojewodztwo/preceding-sibling::*|#*" /> <!-- Copy nodes before 'Wojewodztwo' -->
<Lokal_test />
<xsl:apply-templates select="Wojewodztwo|Wojewodztwo/following-sibling::*|#*" /> <!-- Copy nodes after 'Wojewodztwo' (including) -->
</xsl:copy>
</xsl:template>
The second template puts the Lokal_test element before the Wojewodztwo element and copies the surrounding nodes.

XSL copy without values is it possible?

I want to compare two xmls.
1. First compare XML strucutre/schema.
2. Compare values.
I am using beyond compare tool to compare. Since these two xmls are different values, there are lot many differences in comparison report, for which I am not interested. Since, my focus now is to only compare structure/schema.
I tried to copy the xmls by following template, and other as well. But every time it is with values.
I surfed on google, xsl-copy command itself copies everything for selected node/element..
Is there any ways with which I can filter out values and only schema is copied ?
My Data :
<root>
<Child1>xxxx</Child1>
<Child2>yyy</Child2>
<Child3>
<GrandChild1>dddd<GrandChild1>
<GrandChild2>erer<GrandChild2>
</Child3>
</root>
Template used :
<xsl:template match="/">
<xsl:apply-templates/>
</xsl:template>
<!-- for all elements (tags) -->
<xsl:template match="*">
<!-- create a copy of the tag (without attributes and children) in the output -->
<xsl:copy>
<!-- For all attributes of the current tag -->
<xsl:for-each select="#*">
<xsl:sort select="name( . )" order="ascending" case-order="lower-first" />
<xsl:copy/>
</xsl:for-each>
<!-- recurse through all child tags -->
<xsl:apply-templates/>
</xsl:copy>
</xsl:template>
<xsl:template match="text()|comment()|processing-instruction()">
<xsl:copy/>
</xsl:template>
OutPut Required :
Something like..
<root>
<Child1></Child1>
<Child2></Child2>
<Child3>
<GrandChild1><GrandChild1>
<GrandChild2><GrandChild2>
</Child3>
</root>
At the moment, you have a template matching text() to copy it. What you need to do is remove this match from that template, and have a separate template match, that matches only non-whitespace text, and remove it.
<xsl:template match="comment()|processing-instruction()">
<xsl:copy/>
</xsl:template>
<xsl:template match="text()[normalize-space()]" />
For white-space only text (as used in indentation), these will be matched by XSLT'S built-in templates.
For attributes, use xsl:attribute to create a new attribute, without a value, rather than using xsl:copy which will copy the whole attribute.
<xsl:attribute name="{name()}" />
Note the use of Attribute Value Templates (the curly braces) to indicate the expression is to be evaluated to get the string to use.
Try this XSLT
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
<!-- for all elements (tags) -->
<xsl:template match="*">
<!-- create a copy of the tag (without attributes and children) in the output -->
<xsl:copy>
<!-- For all attributes of the current tag -->
<xsl:for-each select="#*">
<xsl:sort select="name( . )" order="ascending" case-order="lower-first" />
<xsl:attribute name="{name()}" />
</xsl:for-each>
<!-- recurse through all child tags -->
<xsl:apply-templates/>
</xsl:copy>
</xsl:template>
<xsl:template match="comment()|processing-instruction()">
<xsl:copy/>
</xsl:template>
<xsl:template match="text()[normalize-space()]" />
</xsl:stylesheet>
Also note that attributes are considered to be unordered in XML, so although you have code to sort the attributes, and they probably will appear in the right order, you can't guarantee it.

Keep schemaversion while doing <xsl:apply-templates select="#*|node()"/>

I have requirements to copy whole node including schemaversion and namespaces. (Below is part of XML)
<ns0:Body>
<std:test com:schemaVersion = "4.1" xmlns:com = "http://www.test.com/DI/D2P/Broker/Schemas/CommonTypes" xmlns:std = "http://www.test1.com/DI/D2P/Broker/Schemas/test" xmlns:xsi = "http://www.w3.org/2001/XMLSchema-instance">
<com:SystemMessageHeader>
...
...
I am using below xslt to do this task. (i have mentioned only needed portion of XSLT.) With this XSLT, it does copy of whole node including namespaces but it does not copy com:schemaVersion = "4.1". Please help me how can i fix it ?
<nos:Body>
<xsl:for-each select ="nos:GRRCall/nos:Response/ns0:Body/*">
<xsl:copy>
<xsl:variable name="v1" select="position()"/>
<xsl:variable name="v2" select="count(../../../../*)"/>
<xsl:apply-templates />
</xsl:copy>
</xsl:for-each>
</nos:Body>
<xsl:template match="#*|node()">
<xsl:copy>
<xsl:apply-templates select="#*|node()"/>
<xsl:apply-templates select="com:SystemMessageHeader"/>
</xsl:copy>
</xsl:template>
When you perform an apply-templates without specifying a select expression -- as in your stylesheet fragment -- the attributes of the context node are not selected for transformation. That's why the canonical identity transform explicitly selects attributes:
<!-- Identity transform -->
<xsl:template match="node()|#*">
<xsl:copy>
<!-- HERE: -->
<xsl:apply-templates select="node()|#* />
</xsl:copy>
</xsl:template>
Your first alternative, then, would be to follow the model of the identity transform when you apply templates to the children of <ns0:Body> elements.
Your second alternative would be to explicitly transform the wanted attribute by adding ...
<xsl:apply-templates select="#com:schemaVersion" />
... which will not affect other attributes, and will do nothing if the source element has no such attribute.
A third alternative would be to ensure that the transformed elements have the wanted attribute, even if the source elements didn't, by inserting code to create one into the template:
<xsl:attribute name="com:schemaVersion">
<xsl:value-of select="#com:schemaVersion" />
</xsl:attribute>
That particular version will provide an empty string as the attribute value when the source node has no such attribute; you could modify it to provide a default value instead if desired.

how to exclude footer element content from body using xslt 1.0

I have a html page for an example like below
<html><head><title>test</title></head><body><div>test1</div><footer><div>test2</div></footer></body></html>
I have written xslt 1.0 to transform and extract the title and body content, but my requirement is to ignore footer content alone and consider all other element values inside body content. How to achieve this ?
<xsl:template match="/">
<document >
<xsl:copy-of select="#*" />
<xsl:apply-templates select="html/head" />
<xsl:apply-templates select="html/body" />
</document>
</xsl:template>
<xsl:template match="html/head">
<content name="title">
<xsl:value-of select="title" />
</content>
</xsl:template>
<xsl:template match="html/body">
<content name="snippet">
<xsl:value-of select="viv:replace(viv:replace(.,'<[^>]*>',' ', 'gi'),'&nbsp;','','gi')"/>
</content>
</xsl:template>
Q: how to exclude footer element content from body using xslt 1.0
If this is really your question this should be answered hundred times.
Stat with an identity transform and have empty templates for elements to ignore.
<!-- identity transform -->
<xsl:template match="#*|node()">
<xsl:copy>
<xsl:apply-templates select="#*|node()"/>
</xsl:copy>
</xsl:template>
<xsl:template match="body/footer"/>
Looking to your xlst let us assume there a some strange_other_things requested.
Do strange_other_things for the body without footer put the result form identity transfer into a variable.
<xsl:template match="body" mode="strange_other_things">
<xsl:variable name="body" >
<xsl:copy>
<xsl:apply-templates select="#*|node()"/>
</xsl:copy>
</xsl:variable>
<!-- use $body but I'm out here -->
</xsl:template>
Further guess: With viv:replace(.,'<[^>]*>',' ', 'gi') you try to remove xml element names. This will not work because . is used in a text context an will only return all text inside the current node.
So if I'm right the question is a lite deceptive.

orbeon: applying xslt transform in xbl on bound document

In Orbeon Forms I need to create a component (using XBL) that when bound to an instance like
<OCinstructionitem>
<OCp>paragraph 1</OCp>
<OCp>paragraph 2 <OCem>with italics part</OCem> rest of paragraph 2 </OCp>
</OCinstructionitem>
creates an editable div like this:
<div contentEditable="true">
<p>paragraph 1</p>
<p>paragraph 2 <i>with italics part</i> rest of paragraph 2 </p>
</div>
My thought was that I need to do this using XSLT. I get this working when the to-be-transformed XML is inside the xforms document:
<oc:instructionitem>
<OCinstructionitem>
<!-- here the xml of above -->
...
</OCinstructionitem>
</oc:instructionitem>
but I want to let the XSLT operate on the bound node as in:
<oc:instructionitem ref="OCinstructionitem"/>
However, I canot access the bound node from the XSLT.
My question: is that just not possible? Or do I have to modify my XBL?
My XBL code:
<xbl:binding id="oc-instructionitem" element="oc|instructionitem">
<xbl:template xxbl:transform="oxf:xslt">
<xsl:transform version="2.0">
<xsl:template match="#*|node()" priority="-100">
<xsl:apply-templates select="#*|node()"/>
</xsl:template>
<xsl:template match="text()" priority="-100" mode="in-paragraph">
<xsl:copy>
<xsl:value-of select="."/>
</xsl:copy>
</xsl:template>
<xsl:template match="OCem" mode="in-paragraph">
<x:i><xsl:value-of select="."/></x:i>
</xsl:template>
<xsl:template match="OCp">
<x:div>
<xsl:apply-templates mode="in-paragraph"/>
</x:div>
</xsl:template>
<xsl:template match="/*">
<x:div contentEditable="true">
<xsl:apply-templates/>
</x:div>
</xsl:template>
</xsl:transform>
</xbl:template>
</xbl:binding>
</xbl:xbl>
Any help would be greatly appreciated,
edit: after helpfull comment of tohuwawohu (below)
It seems that you need to define a variable which is bound to the instance data. Like this:
<xsl:template match="oc:instructionitem">
<xforms:group xbl:attr="model context ref bind" xxbl:scope="outer">
<xxforms:variable name="binding" as="node()?" xxbl:scope="inner" >
<xxforms:sequence select="." xxbl:scope="outer"/>
</xxforms:variable>
<xforms:group xxbl:scope="inner">
<!-- Variable pointing to external single-node binding -->
<xforms:group ref="$binding">
<xsl:call-template name="main"/>
</xforms:group>
</xforms:group>
</xforms:group>
</xsl:template>
<xsl:template name="main">
<x:div contentEditable="true">
<xforms:repeat nodeset="*">
<xsl:call-template name="paragraph"/>
</xforms:repeat>
</x:div>
</xsl:template>
However, the XSLT elements still cannot act on the data. It can only generate XFORMS elements, which can act on the data. This means that something like <xsl:template match="OCp"> will never be selected. This is why the above code uses named templates.
So the question still stands: can the bound data be made available to the xslt code?
Martijn
I didn't test it, but i think it should be possible by breaking the encapsulation. Without this, your XBL will only have access to the content of the bound node, as in your third code snippet. If you want to make the XBL access the XForms instance data, you will have to declare a variable inside the XBL that's pointing to the single-node binding of the bound node.
Here's the code snippet from the Wiki example:
<xforms:group xbl:attr="model context ref bind" xxbl:scope="outer">
<xforms:group xxbl:scope="inner">
<!-- Variable pointing to external single-node binding -->
<xxforms:variable name="binding" as="node()?">
<xxforms:sequence select="." xxbl:scope="outer"/>
</xxforms:variable>
...
</xforms:group>
</xforms:group>
Having defined the single-node binding like this, it should be possible to reference that variable and use it for xslt transformation. Just replace the XSLT template element matching the root node and point it to the content of the variable $binding:
<xsl:transform version="2.0">
<xsl:template match="#*|node()" priority="-100">
<xsl:apply-templates select="#*|node()"/>
</xsl:template>
<xsl:template match="text()" priority="-100" mode="in-paragraph">
<xsl:copy>
<xsl:value-of select="."/>
</xsl:copy>
</xsl:template>
<xsl:template match="OCem" mode="in-paragraph">
<x:i><xsl:value-of select="."/></x:i>
</xsl:template>
<xsl:template match="OCp">
<x:div>
<xsl:apply-templates mode="in-paragraph"/>
</x:div>
</xsl:template>
<xsl:template match="oc:instructionitem">
<xforms:group ref="$binding">
<x:div contentEditable="true">
<xsl:apply-templates/>
</x:div>
</xforms:group>
</xsl:template>
Hope this works for you. Maybe this requires making provisions if the bound node (in your example: oc:instructionitem) isn't empty, so the xslt may process both that content as well as the content of the $binding variable.