近期有个任务要求处理大XML文件，其中有个存了Base64的大节点（>90M，路径已知）。
这种任务只能上XmlReader，即使如此大节点的处理还是头疼了一阵……
最初查MSDN的时候，找到了ReadChars()，可以拿来对付大节点。
方法说明：https://msdn.microsoft.com/zh-cn/library/system.xml.xmltextreader.readchars(v=vs.110).aspx
示例中提到使用方法是：
while( != reader.ReadChars(buffer, , ))
{
// Do something.
// Attribute values are not available at this point.
}
这个处理规范格式的XML没有问题，比如这样的：
<Root>
<LeafNode>Value</LeafNode>
<ParentNode>
<LeafNode>Value</LeafNode>
</ParentNode>
</Root>
但是（没人喜欢这个词，然并卵……），遇到些格式诡异的XML就……
<Root><LeafNode>Value</LeafNode><ParentNode>
<LeafNode>Value</LeafNode></ParentNode>
</Root>
比如这个画风的，用示例代码去读第一个LeafNode的内容，估计会读出“ValueValue”来……
偏偏输入的XML就是这风格的……（*sigh*）
单步执行了一阵，发现这种情况下，XmlTextReader.Name会变化成下个节点的名称（XmlTextReader.LocalName亦如此），可以根据这个判断是否已经达到节点结尾。
改进版为：
string currentName = reader.LocalName;
while(currentName == reader.LocalName && != reader.ReadChars(buffer, , ))
{
// Do something.
// Attribute values are not available at this point.
}
顺便贴上一个转写并对特定节点进行处理的代码：
List<string> processNodePathList = new List<string> {"/Root/Path/to/Target"};
List<string> bigNodePathList = new List<string> { "/Root/Path/to/Big/Node" }; private static void ProcessBigXmlFile(string sourcePath, string targetPath, IList<string> processNodePathList, IList<string> bigNodePathList)
{
var processNodeNameList =
processNodePathList.Select(
processNodePath => processNodePath.Split(new[] { '/' }, StringSplitOptions.RemoveEmptyEntries))
.Select(nodePathParts => nodePathParts[nodePathParts.Length - ])
.ToList();
var bigNodeNameList = bigNodePathList.Select(
bigNodePath => bigNodePath.Split(new[] { '/' }, StringSplitOptions.RemoveEmptyEntries))
.Select(nodePathParts => nodePathParts[nodePathParts.Length - ])
.ToList(); var sourceStream = new FileStream(sourcePath, FileMode.Open, FileAccess.Read);
var reader = new XmlTextReader(sourceStream); var targetStream = new FileStream(targetPath, FileMode.Create, FileAccess.Write);
var writer = new XmlTextWriter(targetStream, Encoding.UTF8); try
{
var pathStack = new Stack<string>();
var readResult = reader.Read();
while (readResult)
{
int skipMode = ;
switch (reader.NodeType)
{
case XmlNodeType.Element:
{
pathStack.Push(reader.Name);
writer.WriteStartElement(reader.LocalName);
if (reader.HasAttributes)
{
while (reader.MoveToNextAttribute())
{
writer.WriteAttributeString(reader.LocalName,
reader.Value);
}
reader.MoveToElement();
} if (processNodeNameList.Contains(reader.LocalName))
{
var index = processNodeNameList.IndexOf(reader.LocalName);
if (CompareNodePath(pathStack, processNodePathList[index]))
{ // Replace node content writer.WriteFullEndElement();
skipMode = ;
}
}
else if (bigNodeNameList.Contains(reader.LocalName))
{
var index = bigNodeNameList.IndexOf(reader.LocalName);
if (CompareNodePath(pathStack, bigNodePathList[index]))
{
reader.MoveToContent();
var buffer = new char[];
int len;
while (reader.LocalName == bigNodePathList[index] &&
(len = reader.ReadChars(buffer, , buffer.Length)) > )
{
writer.WriteRaw(buffer, , len);
}
writer.WriteFullEndElement();
skipMode = ;
}
}
if (reader.IsEmptyElement)
{
pathStack.Pop();
writer.WriteEndElement();
}
break;
}
//case XmlNodeType.Attribute:
//{
// newPackageWriter.WriteAttributeString(oldPackageReader.LocalName, oldPackageReader.Value);
// break;
//}
case XmlNodeType.Text:
{
writer.WriteValue(reader.Value);
break;
}
case XmlNodeType.CDATA:
{
writer.WriteCData(reader.Value);
break;
}
//case XmlNodeType.EntityReference:
//{
// newPackageWriter.WriteEntityRef(oldPackageReader.Name);
// break;
//}
//case XmlNodeType.Entity:
//{
// break;
//}
case XmlNodeType.ProcessingInstruction:
{
writer.WriteProcessingInstruction(reader.Name, reader.Value);
break;
}
case XmlNodeType.Comment:
{
writer.WriteComment(reader.Value);
break;
}
//case XmlNodeType.Document:
//{
// break;
//}
case XmlNodeType.DocumentType:
{
writer.WriteRaw(string.Format("<!DOCTYPE{0} [{1}]>", reader.Name,
reader.Value));
break;
}
//case XmlNodeType.DocumentFragment:
//{
// break;
//}
//case XmlNodeType.Notation:
//{
// break;
//}
case XmlNodeType.Whitespace:
{
writer.WriteWhitespace(reader.Value);
break;
}
//case XmlNodeType.SignificantWhitespace:
//{
// break;
//}
case XmlNodeType.EndElement:
{
pathStack.Pop();
writer.WriteFullEndElement();
break;
}
case XmlNodeType.XmlDeclaration:
{
writer.WriteStartDocument();
break;
}
} switch (skipMode)
{
case :
{
reader.Skip();
pathStack.Pop();
readResult = !reader.EOF;
break;
}
case :
{
pathStack.Pop();
readResult = !reader.EOF;
break;
}
default:
{
readResult = reader.Read();
break;
}
}
}
}
finally
{
writer.Close();
targetStream.Close();
targetStream.Dispose();
reader.Close();
sourceStream.Close();
sourceStream.Dispose();
}
}private static bool CompareNodePath(Stack<string> currentNodePathStack, string compareNodePathString)
{
var currentArray = currentNodePathStack.Reverse().ToArray();
var compareArray = compareNodePathString.Split(new[] { '/' }, StringSplitOptions.RemoveEmptyEntries);
if (compareArray.Length != currentArray.Length)
{
return false;
}
bool isDifferent = false;
for (int i = ; i < currentArray.Length; i++)
{
if (compareArray[i] != currentArray[i])
{
isDifferent = true;
break;
}
}
return !isDifferent;
}
