我需要将文件的前100个字节移动到文件的末尾,然后保存该文件(windows窗体应用程序)。然后我需要反向过程(将结尾处的100个字节移回到开头然后再次保存)。
其中一些文件非常大(超过2GB),因此我无法使用file.readallbytes,因为我的内存不足。
我尝试过使用文件流并使用filestream.position,但我无法绕过移动字节然后保存文件。
任何指导都将不胜感激。
答案 0 :(得分:4)
您不应将所有数据读入内存。通过使用FileStream,您可以读取多个数据块,例如每个1 KB,并将其存储在新文件中。从位置100开始,跳过第一个字节。重新排列完整文件后,在末尾添加跳过的字节。最后将新文件移动到旧文件的位置。通过提高maxBufferSize,您可以加快复制过程,但需要使用更多内存。
要还原更改,请从最后100个字节开始,然后从开始直到inputStream.Length - 100
继续。
string inputFile = "C:\\input.txt";
string tempFile = "C:\\input.txt";
int dataLength = 100;
int maxBufferSize = 1024;
using (var inputStream = new FileStream(inputFile, FileMode.Open, FileAccess.Read)) {
int length = (int)inputStream.Length;
int currentInputPosition = dataLength;
inputStream.Position = currentInputPosition;
using (var outputStream = new FileStream(tempFile, FileMode.Create, FileAccess.ReadWrite)) {
var bufferSize = Math.Min(maxBufferSize, length - currentInputPosition);
var buffer = new byte[bufferSize];
while (inputStream.Read(buffer, 0, bufferSize) > 0) {
currentInputPosition += bufferSize;
outputStream.Write(buffer, 0, buffer.Length);
}
buffer = new byte[dataLength];
inputStream.Position = 0;
inputStream.Read(buffer, 0, buffer.Length);
outputStream.Write(buffer, 0, buffer.Length);
}
}
File.Delete(inputFile);
File.Move(tempFile, inputFile);
答案 1 :(得分:3)
关键是使用字节缓冲区,因此您一次只处理文件的一小部分。听起来你有正确的方法,但是沿途还有一些问题。
这是一个展示我如何接近它的例子:
public enum SwapType
{
FrontToBack,
BackToFront
}
public static class EndSwap
{
public static void DoSwap(string path, SwapType swapType)
{
if (path == null)
{
throw new ArgumentNullException("path",
"You must supply a path to the file.");
}
if (!File.Exists(path))
{
throw new FileNotFoundException("File not found.");
}
string tempPath = Path.GetTempFileName();
byte[] buffer = new byte[4096];
byte[] swapBytes = new byte[100];
using (FileStream inputFs = new FileStream(path, FileMode.Open,
FileAccess.Read))
using (FileStream outputFs = new FileStream(tempPath,
FileMode.Open, FileAccess.Write))
{
int bytesRead = -1;
if (swapType == SwapType.FrontToBack)
{
// We want to keep hold of the first 100 bytes of the file
// and output them after copying the rest of file
inputFs.Read(swapBytes, 0, 100);
}
else
{
// Read the last 100 bytes of the file
inputFs.Seek(-100, SeekOrigin.End);
inputFs.Read(swapBytes, 0, 100);
// Output them straight to the output file
outputFs.Write(swapBytes, 0, 100);
// Reposition to the beginning of the input file
inputFs.Seek(0, SeekOrigin.Begin);
}
// The number of bytes left to copy is 100 less than the file
// length
long bytesRemaining = inputFs.Length - 100;
// Copy the rest of the bytes
while (bytesRemaining > 0)
{
bytesRead = inputFs.Read(buffer, 0, buffer.Length);
// NB: the number of bytes read could be more than the
// number remaining
outputFs.Write(buffer, 0,
(int)Math.Min(bytesRead, bytesRemaining));
bytesRemaining -= bytesRead;
}
// Don't forget to append the start bytes if required
if (swapType == SwapType.FrontToBack)
{
outputFs.Write(swapBytes, 0, 100);
}
}
// Now swap the files themselves
File.Delete(path);
File.Move(tempPath, path);
// NB: could do File.Replace() if backup is needed
}
}
使用示例:
// Copy first 100 bytes to end
EndSwap.DoSwap(@"C:\Users\Dave\Downloads\MyTest.pdf", SwapType.FrontToBack);
// Copy those same 100 bytes back to the beginning again
EndSwap.DoSwap(@"C:\Users\Dave\Downloads\MyTest.pdf", SwapType.BackToFront);
当然,最终结果与原始文件相同。