Re: How to preserve numeric entities when converting xml String to a org.w3c.dom.Document ?

Artur Tomusiak schrieb:
>
> I am trying to convert a String with XML content in it into the
> org.w3c.dom.Document object to do some modifications and then to
> convert it back to the String. However, even if I do not do any
> modifications to the object, I am still getting back a different
> String than what I have provided as an input. The problem is with
> the numeric XML entities. For example, if my input String is:
>
> <?xml version="1.0" encoding="UTF-8"?>
> <xml>
>    &#169;
>    &#38;   </xml>

Hi Artur,

in fact, and to be pedantic, these are neither entities nor entity
references; they're numerical character references; they just happen
to use the same syntax as general entity references. (See XML spec
if interested.)

As keshlam said, these are 100 % identical as far as XML is concerned.

It's not clear to me whether you use XSLT at all or only the DOM.
I'm assuming you're using XSLT.

When transforming to a DOM target, the XSLT serialization instruction
like <xsl:output encoding="US-ASCII"/> is disregarded.

If all you want is a string, there is no point in transforming to the
DOM. In that case, simply specify <xsl:output encoding="US-ASCII"/> in
your stylesheet. That would force numerical character references for
non-ASCII characters.

But the characters in your example are ASCII characters, and I do not
know of a way to have them serialized as numerical character references
in XSLT 1.0. Use Perl or AWK or some other general text processing tool
to postprocess your output.

Michael Ludwig