This is the final version sent on 14-9-04

http://in.geocities.com/rsramsam/xslt.htm

XSLT in J2EE/JAX & ASP.Net

(published in DeveloperIQ..October-2004)

R.S.RAMASWAMY

(rsramsam@yahoo.co.in)

This is a 4 part tutorial

XSLT BASICS & DEMOS IN MICROSOFT BROWSER .PART-1(HERE)

XSLT IN DOTNET PLATFORM. .PART-2

XSLT PROCESSING USING XALAN IN JAVA .PART-3

XSLT IN JSP USING BEAN .PART-4

PART-1

------

XSLT BASICS & DEMOS IN MICROSOFT BROWSER

=======================================

In the previous instalments(july & august ,2004), we had seen how jdk1.4 simplifies XML processing in java by incorporating JAXP1.1 in Jdk itself.

We saw how XML processing can be done by using DOM , SAX and JDOM. However, there is yet another method known as XSLT (Extensible Style-sheet Transformation).

Xalan is the most well-known XSLT processor and is now available from Apache Software Foundation.This is the default XSLT processor in jdk1.4.

In this part of the tutorial, we will acquaint ourselves with XSL terminology and techniques , a standard maintained by W3C (WorldWideWeb Consortium).

XML technologies like DOM,SAX and XSLT are important not only for J2EE programmers but ASP.net and Perl/PHP/Python programmers as well.

In fact, the entire DotNet framework has been built with easy XML transformations in mind, from the ground-up.Microsoft has been one of the main driving forces behind the adoption of XML and had the foresight to incorporate very powerful support for XML and XSLT in Internet-Explorer browser, long back.Netscape followed suit only later.Someone lamented in the web forum that the OSF community 'won the battle but lost the war' by conceding leadership to Microsoft in browser technologies.Ultimately it is the mind-share of the end-user that decides the issue . SOAP (Simple Object Access Protocol) also is a standard , proposed and promoted by Microsoft, though it has now become a public standard.

( DON BOX of Developmentor and Dr.Gopal Kakivaya of Microsoft labs were the lead authors of SOAP). As mentioned in the previous instalment of this tutorial, this could even lead us sometimes to identify XML too closely with ready-made xml-processing in Microsoft browser (client-side).Fortunately for open-source adherents, Mozilla browser , has caught up with Internet Explorer in its xslt capabilities.Netscape6 also offers similar facilities now.

[However, the advent of server-side xsl transformation, has made browser-side xslt 'not-so-vital', now.]

Jon Bosak of Sun Micro-system was one of the lead- persons in the creation of XML specifications. And IBM has done its own share by donating Lotus-XSLT Processor to Apache Foundation.(Xalan is based on this).Thus, all the leading corporates have played their role in developing XML standards and tools.Individuals like Michael Kay (SAXON xslt processor) also have contributed.As XML-WebService unifies all languages and platforms, it will be short-sighted if we persist in old-ways of thinking, pitting one company against another, though each has its own merits and demerits.We will adopt each technology by its suitability for a given requirement , rather than by its source of creation.’Horses for the Courses! ‘, as they say.

Granting that XML is 'platform and language neutral data' and Java is 'platform-neutral language' ,the third leg of the tripod is missing. How do we translate XML data into other forms?That is where XSLT, appears.This potential for transformation is what sets XML apart from other data-formats like HTML, PDF and PostScript.

Web applications must support a number of different types of clients such as HTML browsers, WAP browsers and other types for wireless clients. It is not realistic to create the data specifically for each client. The better method is to have a single XML file but distribute this to different clients , after suitable 'transformation'. ( that explains why we are speaking of TrAX ..' Transformation API for XML').

Looking at a possible scenario in j2ee environment, the business-logic is encapsulated in an EJBean. The EJBean is called by either a servlet or jsp.

The data returned by the EJBean is converted into XML form.Ready-made software (freeware ,even) is available for this purpose. Even otherwise, it can be done by us programmatically . It is always easy to convert a relational table into its xml equivalent . But , to convert an xml file in hierarchial data form into a relational table is admittedly difficult and clumsy.Luckily for programmers, Enterprise data so far is mainly in Relational databases.

Once the query result is available in xml form, it can be sent for display in different formats,suitable for standard web-browsers and mobile clients. For this purpose, we can use different style-sheets.(XSL stands for Extensible style-sheet Language).

XSL is split up into two different specifications.

a) XSL-FO ( Formatting Objects)

b) XSLT (Transformation)

XSL-FO is purely about presentation while XSLT is presentation & transformation. XSLT is considered by many experts as the most important part of XML initiative.And 'Cocoon', a famous open-source & free software from Apache is entirely based on XSLT.

It is customary to speak of clear separation of business-logic and presentation. Thus, in a web application, JSP is recommended ,rather than a pure servlet, because it is easier to encapsulate the business-logic in a jsp-bean and minimize the code in a dynamically-created page, leaving the job of suitable presentation to page-designers, who need not know Java language and programming atall.Such division of labour , is seen to improve the quality of the result and also make maintenance easier.

----

XML & XSLT combination goes one step further. Here, we have clean separation between logic, data and presentation.

a) Ejbean deals with the logic

b) data is in XML

c) presentation is taken care of, by XSLT.

Is XSLT then all about transformation for presentation only? Not atall!XSLT can be used like SQL. We can use xslt to query, sort and modify,the xml data itself.Thus it is not a mere presentation technology but a data-extraction and data-manipulation technology as well.

Actually, XSLT is a PROGRAMMING LANGUAGE concealed and expressed in XML syntax. Among programming languages, there are many types.Unlike languages like C, C++, Java , Pascal etc which are known as 'Imperative Languages', XSLT belongs to the 'functional languages' group.Its approach has some resemblance to languages like Lisp/ Scheme. As it is expressed in familiar html/xml syntax , it is deceptively simple. From another perspective, we can think of XSLT as being similar to SQL of database queries, where we simply state what we want and do not dictate how it should be done.

Let us now consider a simple example. We are having an XML file named 'students.xml'.

( please remember , that XML is case-sensitive).Just to get familiar with the idea, we will consider displaying this xml file after transforming it by using a style-sheet. We are using Internet Explorer-6 and using its built-in xslt capabilities.

[ The emphasis in this tutorial is not so much on in-depth treatment of XSLT but on implementing it in J2EE & ASP.net environments.]

The students.xml file is given below.

***********************************************

// students.xml

<?xml version="1.0"?>

<name>Thomas</name>

<place>Delhi</place>

</student>

<name>David</name>

<place>Bombay</place>

</student>

<name>Mathew</name>

<place>Bangalore</place>

</student>

<place>Madras</place>

</student>

</students>

---------------------------------------------------

The following style-sheet is used for

displaying the above xml file.

// xsl1.xsl

==============

<?xml version="1.0"?>

<xsl:stylesheet

xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

<xsl:template match="/">

<html>

<body>

<tr>

<th>Place</th>

<th>Number</th>

</tr>

<xsl:for-each select="students/student">

<tr>

</tr>

</xsl:for-each>

</table>

</body>

</html>

</xsl:template>

</xsl:stylesheet>

We are having students.xml and xsl1.xsl in ‘C:\xslt’ folder. let us now view ‘c:\xslt\students.xml’ in InternetExplorer-6.We will get a display of the file in xml format with tree structure,just as the original xml file.Let us now make a reference to xsl1.xsl in the students.xml file and rename it as xslstudents.xml

The modified file xslstudents.xml is shown below.

----------------------------------------------------------

// xslstudents.xml

<?xml version="1.0" ?>

<?xml-stylesheet type="text/xsl" href="xsl1.xsl" ?>

<name>Thomas</name>

<place>Delhi</place>

0<number>1111</number>

</student>

// as before

<place>Madras</place>

</student>

</students>

------------------------------------------------

If we view c:\xslt\xslstudents.xml in IE6,

we get the data displayed as html page with table.

Name	Place	Number	Mark
Thomas	Delhi	1111	78
David	Bombay	2222	90
Mathew	Bangalore	3333	92
John	Madras	4444	72

So, we have invoked the xml file with ‘built-in’ style sheet reference and viewed it in IE6.

Just a bit of explanation is in order.

The declaration MUST assign a version number.

The XSLT schema location is at :

'http://www.w3.org/1999/XSL/Transform' & is assigned as namespace. It is purely an internal reference and we don't have to be connected to the internet to use the xsl file which uses this namespace reference!But, we should type this without the slightest mistake , especially regarding case.

-- The for-each tag has now crept into jdk also and so should be familiar to asp.net and jdk coders. It is a kind of looping. We are asking the code to select and display in table cells the values of name, place, number and mark. for each in 'students/student' path.There comes the other piece of the specification, named XPath. It can be thought of as similar to the familiar file hierarchy

-----------------------------------------------------

We will now see four more examples and the resulting display.

xsl2.xsl

<?xml version="1.0"?>

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

<xsl:template match="/">

<html>

<body>

sorting in ascending order by name of the student.

<xsl:for-each select="students/student">

<xsl:sort select="name"

order="ascending" />

<tr>

etc: ( as before )

Result

======

David	Bombay	2222	90
John	Madras	4444	72
Mathew	Bangalore	3333	92
Thomas	Delhi	1111	78

xsl3.xsl

<?xml version="1.0"?>

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

<xsl:template match="/">

SORTING BY name (Descending)

<xsl:for-each select="students/student">

<xsl:sort select="name"

order="descending" />

<tr>

etc...( as before)

result

======

Thomas	Delhi	1111	78
Mathew	Bangalore	3333	92
John	Madras	4444	72
David	Bombay	2222	90

===============================================

xsl4.xsl

<?xml version="1.0"?>

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

<xsl:template match="/">

<html>

<body>

<xsl:for-each select="students/student">

<xsl:sort select="mark"

data-type="number"

order="ascending" />