Transformation Problem For Arabic/French Character : HTML to XML

[Tanzy Akhtar]

Hi,

I have a program, which converts html into xml.
I am facing problem when transforming french/arabic character.

The code is as follows--

    
import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;

import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerException;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.stream.StreamResult;
import javax.xml.transform.stream.StreamSource;

public class XMLTransform 
	{
	
		public static void main(String ags[]) throws FileNotFoundException, TransformerException 
			{      
				  StreamSource source = new StreamSource(new File("D:/Old.xml"));
				  StreamSource stylesource = new StreamSource(new File("D:/Feature_Details.xsl"));
				  TransformerFactory factory = TransformerFactory.newInstance();
				  Transformer transformer = factory.newTransformer(stylesource);
				 
				  FileOutputStream fos =  new FileOutputStream(new File("D:/Transformed.xml"));
				  StreamResult result = new StreamResult(fos);
				  transformer.transform(source, result);
			}
	}

The transformer used is --

com.sun.org.apache.xalan.internal.xsltc.trax.TransformerFactoryImpl

and the version of jar is xalan-2.7.0.

Any idea, what i am missing?

Thanks,
Tanzy.

[Aneesh Vijendran]

Hi,

By default xalan uses UTF-8, so there is no problem with the transformer/parser. The problem should b with your xsl.

Could you try giving the UTF-8 encoding in your xsl. You could do that by adding the following in the xsl file:

<xsl:output
method="xml"
encoding="UTF-8"

....
...
/>

You can have a more look for the syntax here:

http://www.w3schools.com/xsl/el_output.asp


Cheers
Aneesh

[Tanzy Akhtar]

Thank You Aneesh for the respoonse.

<xsl:output
method="xml"
encoding="UTF-8"

....
...
/>

This is been already done.

Meaning, there is some other issue.

Thanks,
Tanzy.

[Aneesh Vijendran]

Could you please attach the xsl & xml here, through the attachments?

Cheers
Aneesh

[Tanzy Akhtar]

Aneesh, this program is running fine if i am not using xalan transformer.

Problem occurs only when using xalan.
Is there any other transformer to which i should use instead of xalan?

[Aneesh Vijendran]

I didnt get the point:

Aneesh, this program is running fine if i am not using xalan transformer

Are you getting

???

in the transformed xml. or any exception. Could you please let me know the exact problem?

You can try using cocoon. But I can say the problem is not with the transformer. I have done numerous transformations with a variety of unicode Indian characters.

Cheers
Aneesh

[Rob Spoor]

tanzy akhtar wrote:I am facing problem when transforming french/arabic character.

What's the problem? http://faq.javaranch.com/java/TellTheDetails

[Tanzy Akhtar]

Are you getting

???

Yes exactly Aneesh. This is my problem. After transformation arabic/french character get replaced by "???".


Sorry Rob, i could not specify my problem earlier and thanks for pointing that.

[Tanzy Akhtar]

Before doing actual transformation, replacing "nbsp" character with "\n".
Below is the program which gets execute before the transformation takes place--

  
String htmlFile = "" + name[i];
final FileInputStream sourceStream = new FileInputStream(new File(htmlFile));
final InputStreamReader isr = new InputStreamReader(sourceStream);
final BufferedReader bufferedReader = new BufferedReader(isr);
final StringBuffer stringBuffer = new StringBuffer();
String str;
while ((str = bufferedReader.readLine()) != null) {
   str = str.replaceAll("nbsp;", "\n");

   if (str.indexOf("!DOCTYPE") < 0) {
	  stringBuffer.append(str);
   }
}

bufferedReader.close();
isr.close();
sourceStream.close();
final FileOutputStream targetStream = new FileOutputStream(htmlFile);
targetStream.write(stringBuffer.toString().getBytes());
targetStream.close();

Here may be some problem when copying.

[Rob Spoor]

&nbsp; is not the same as an enter. The nearest equivalent is a space. That's where the name comes from: non breaking space.

[Tanzy Akhtar]

Thanks Rob.

Well is that nbsp; creating problem in my case?

[Aneesh Vijendran]

Tanzy,

Shoudn't it be

 rather than [b]nbsp;[/b] ?

Well what I think is, there isn't nay problem with your code, I guess it's rather the problem with your browser. Did you try making the browser charset to unicode?

(Mozilla) View ->Character Encoding -> Unicode
(IE) View ->Enclding->Unicode

Let me know this.

Cheers
Aneesh

[Rob Spoor]

tanzy akhtar wrote:Thanks Rob.

Well is that nbsp; creating problem in my case?

Possibly;   is a valid HTML entity, but unless you declare it explicitly again, it's not in XML. For instance, the following XML document gives me the following errors when running through xmllint:

<root>&nbsp;</root>

file:///D%3A/test.xml:1: parser error : Entity 'nbsp' not defined
<root>&nbsp;</root>
            ^

[Aneesh Vijendran]

Rob, I guess it's becasue &amb;nbsp; is not an xml entity, it's only an html entity. So I guess his problem seems to be the encoding issue.

Cheers
Aneesh

[Tanzy Akhtar]

Shoudn't it be

 rather than nbsp; ?

Yes Aneesh. It is the same character you are saying.
Actually while posting the ampersand character, it got invisible.

Thanks for correcting this.

Well, browser setting is also done, even though it's not working.

[Aneesh Vijendran]

hi,

Have send you a Pm with my email.


Cheers
Aneesh

[Rob Spoor]

Aneesh Vijendran wrote:Have send you a Pm with my email.

http://faq.javaranch.com/java/UseTheForumNotEmail
Don't use email or private messages to come to a solution; other people will not see those so you are withholding that solution from everybody else.

[Aneesh Vijendran]

Rob, I would definitely post the solution here, if ever I come to a solution. If there are some files which he can't put coderanch, what could be done?

Cheers
Aneesh

[Rob Spoor]

He should obfuscate them - replace all sensitive data. I once had to export our customer database; I ended up giving them all my manager's name in the export

[Aneesh Vijendran]

yeah you are right

oops manager's details lol!. He might have got a hundred calls regarding market research and stuff lol!

[Tanzy Akhtar]

lolz...
nice conversation..

[Tanzy Akhtar]

Hi,

I got work around of my problem..

Just put "UTF-8" as parameter wherever creation of inputstream/outputstream takes place.

It's working fine for me.

Thank you Rob and Aneesh for useful guidelines.

Life Rocks,
Tanzy.

[Aneesh Vijendran]

Excellant!!!