XOM API for Processing XML with Java

["Michael Kay" <mike AT saxonica.com>]

Your XML examples aren't well-formed, which makes it difficult to help you
as we can't run your test cases.

Whitespace is significant: if you want to eliminate it from the comparison,
you'll have to do this by hand (e.g. by a transformation applied before
canonicalizing).

Attribute order, and the difference between <a></a> and <a/>, should not be
significant.

Order of elements is significant.

You might like to look at the saxon:deep-equal() function as an alternative
approach. 

http://www.saxonica.com/documentation/extensions/functions/deepequal.html

It takes a set of flags to define what features of the XML you consider
significant or insignificant, and there is also an option to get
explanations if the documents are deemed non-equal (though that doesn't seem
to be needed for your application). It can be called from XPath and is also
available directly from Java as the static method
net.sf.saxon.functions.DeepEqual.deepEquals().

Michael Kay
http://www.saxonica.com/ 

> -----Original Message-----
> From: xom-interest-bounces AT lists.ibiblio.org 
> [mailto:xom-interest-bounces AT lists.ibiblio.org] On Behalf Of 
> Marcelo Schneider
> Sent: 10 October 2008 12:59
> To: xom-interest AT lists.ibiblio.org
> Subject: [XOM-interest] Comparing two semantically equal XML files
> 
> Good morning list.
> 
> In our company we currently store some graphs in XML form: 
> the client can create a graph in a Java applet, that is sent 
> to the server and stored as a file. Currently, all the graph 
> <-> XML transformation is done using JDom.
> 
> What I need to do is, when a user opens an already existing 
> graph and tries to save it, compare the new XML sent with the 
> previously existing XML in the server, saving only if they 
> are different (so changing dates are correct).
> 
> Googling around for a while I found out that a possible way 
> to do that without worrying with XML structure was to use 
> Canonicalization, and that XOM could possibly help me with 
> that. Is this assumption correct? 
> Here is what I'm currently doing:
> 
> | public boolean areSemanticallyEqual(String oldXML, String newXML) {
> |     Builder builder = new Builder();
> |     Document oldDoc = builder.build(oldXML);
> |     Document newDoc = builder.build(newXML);
> |     // here canonicalize the old xml
> |     ByteArrayOutputStream oldStream = new ByteArrayOutputStream();
> |     Canonicalizer outputter = new Canonicalizer(oldStream , false);
> |     outputter.write(oldDoc);
> |     // here canonicalize the new xml
> |     ByteArrayOutputStream newStream = new ByteArrayOutputStream();
> |     Canonicalizer outputter = new Canonicalizer(newStream, false);
> |     outputter.write(newDoc);
> |     return Arrays.equals(oldStream.toByteArray(),
> newStream.toByteArray());
> | }
> 
> However, the above code returns false when I compare XMLs 
> like these (they all start with "<?xml version="1.0" 
> encoding="UTF-8"?>"):
> 
> xml1: "<tag1><tag2 id="1" name="a">Some 
> Text</tag2></tag1><other tag />"
> xml2: "<tag1>\n\t<tag2 id="1" name="a">Some 
> Text</tag2>\r</tag1><other tag />" (with new lines, tabs and 
> carriage returns)
> xml3: "<tag1><tag2 id="1" name="a">Some Text</tag2></tag1><other 
> tag></other tag>" (other valid xml representation)
> xml4: "<tag1><tag2 name="a" id="1">Some 
> Text</tag2></tag1><other tag />" 
> (attributes order reversed)
> xml5: "<other tag /><tag1><tag2 id="1" name="a">Some 
> Text</tag2></tag1>" 
> (yet another thing that looks the same, but I'm probably 
> wrong assuming this should count as equal)
> 
> I would expect the return to be always true, except for 
> (maybe?) xml1 > xml5. Is this correct? Any hints are highly 
> appreciated (:
> 
> --
> Marcelo Frantz Schneider
> SIC - TCO - Tecnologia em Engenharia do Conhecimento DÍGITRO 
> TECNOLOGIA
> 
> 
> 
> -- 
> Esta mensagem foi verificada pelo sistema de antivírus da Dígitro e
> acredita-se estar livre de perigo.
> 
> _______________________________________________
> XOM-interest mailing list
> XOM-interest AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/xom-interest