I'm trying to get a newline into a text node using XText
from the Linq XML namespace.
I have a string which contains newline characters however I need to work out how to convert these to entity characters (i.e.
) rather than just having the开发者_高级运维m appear in the XML as new lines.
XElement element = new XElement( "NodeName" );
...
string example = "This is a string\nWith new lines in it\n";
element.Add( new XText( example ) );
The XElement
is then written out using an XmlTextWriter
which results in the file containing the newline rather than an entity replacement.
Has anyone come across this problem and found a solution?
EDIT:
The problem manifests itself when I load the XML into EXCEL which doesn't seem to like the newline character but which accepts the entity replacement. The result is that newlines aren't showing in EXCEL unless I replace them with
Nick.
Cheating:
XmlWriterSettings settings = new XmlWriterSettings();
settings.Indent = true;
settings.CheckCharacters = false;
settings.NewLineChars = " ";
XmlWriter writer = XmlWriter.Create(..., settings);
element.WriteTo(writer);
writer.Flush();
UPDATE:
Complete program
using System;
using System.Xml;
using System.Xml.Linq;
namespace ConsoleApplication1
{
class Program
{
static void Main(string[] args)
{
XElement element = new XElement( "NodeName" );
string example = "This is a string\nWith new lines in it\n";
element.Add( new XText( example ) );
XmlWriterSettings settings = new XmlWriterSettings();
settings.Indent = true;
settings.CheckCharacters = false;
settings.NewLineChars = " ";
XmlWriter writer = XmlWriter.Create(Console.Out, settings);
element.WriteTo(writer);
writer.Flush();
}
}
}
OUTPUT:
C:\Users\...\\ConsoleApplication1\bin\Release>ConsoleApplication1.exe
<?xml version="1.0" encoding="ibm850"?> <NodeName>This is a string With new lines in it </NodeName>
To any standard XML parser there is no difference between the entity
and a new line character, as they are one and the same thing.
To illustrate this the following code shows that they are the same thing:
string s1 = "<root>Test Test2</root>";
string s2 = "<root>Test\nTest2</root>";
XDocument doc1 = XDocument.Parse(s1);
XDocument doc2 = XDocument.Parse(s2);
Console.WriteLine(doc1.ToString());
Console.WriteLine(doc2.ToString());
It's the XmlTextWriter which is responsible for outputting escaped entities. So if you do this, for example:
using (XmlTextWriter w = new XmlTextWriter("test.xml", Encoding.UTf8))
{
w.WriteString("");
}
You will also get an escaped ampersand output in text.xml &#x10
, which you don't want. You would like to keep the 
sequence raw, as is.
The solution I propose is to create a new StreamWriter implementation capable of detecting an escaped string like "&#x10;
":
// A StreamWriter that does not escape characters
public class NonXmlEscapingStreamWriter : StreamWriter
{
private const string AmpToken = "amp";
private int _bufferState = 0; // used to keep state
// add other ctors overloads if needed
public NonXmlEscapingStreamWriter(string path)
: base(path)
{
}
// NOTE this code is based on the assumption that StreamWriter
// only overrides these 4 Write functions, which is true today but could change in the future
// and also on the assumption that the XmlTextWrite writes escaped values in a specific WriteXX calls sequence
public override void Write(char value)
{
if (value == '&')
{
if (_bufferState == 0)
{
_bufferState++;
return; // hold it
}
else
{
_bufferState = 0;
}
}
else if (value == ';')
{
if (_bufferState > 1)
{
_bufferState++;
return;
}
else
{
Write('&'); // release what's been held
Write(AmpToken);
_bufferState = 0;
}
}
else if (value == '\n') // detect non escaped \n
{
base.Write(" ");
return;
}
base.Write(value);
}
public override void Write(string value)
{
if (_bufferState > 0)
{
if (value == AmpToken)
{
_bufferState++;
return; // hold it
}
else
{
Write('&'); // release what's been held
_bufferState = 0;
}
}
base.Write(value);
}
public override void Write(char[] buffer, int index, int count)
{
if (_bufferState > 2)
{
_bufferState = 0;
base.Write('&'); // release this anyway
string replace;
if ((buffer != null) && ((replace = GetReplaceLength(buffer, index, count)) != null))
{
base.Write(replace);
base.Write(buffer, index + replace.Length, count - replace.Length);
return;
}
else
{
base.Write(AmpToken); // release this
base.Write(';'); // release this
}
}
base.Write(buffer, index, count);
}
public override void Write(char[] buffer)
{
Write(buffer, 0, buffer != null ? buffer.Length : 0);
}
private string GetReplaceLength(char[] buffer, int index, int count)
{
// this is specific to the 10 character but could be adapted
const string token = "#10;";
if ((index + count) < token.Length)
return null;
// we test the char array to avoid string allocations
for(int i = 0; i < token.Length; i++)
{
if (buffer[index + i] != token[i])
return null;
}
return token;
}
}
And you can use it like this:
using (XmlTextWriter w = new XmlTextWriter(new NonXmlEscapingStreamWriter("test.xml")))
{
element.WriteTo(w);
}
NOTE: Although it is capable of detecting lonely \n sequences, I suggest you ensure all \n
are actually escaped in your original text, so, you need to replace \n
by 
before you actually output xml, like this:
string example = "This is a stringWith new lines in it";
精彩评论