在C＃中替换文本文件中特定行的最佳方法

Question

For the following operation:

1. Open a text file
2. Search and replace all searching characters with new characters

I'd like to achieve above in c#, here is my code:

using (StreamReader sr = new StreamReader(@"S:\Personal Folders\A\TESTA.txt"))
{
    using (StreamWriter sw = new StreamWriter(@"S:\Personal Folders\A\TESTB.txt"))
    {
        string line;

        while ((line = sr.ReadLine())!= null)
        {               
            if (!line.Contains("before"))
            {
                sw.WriteLine(line);
            }

            else if (line.Contains("before"))
            {
                sw.WriteLine(line.Replace("before", "after"));
            }
        }
    }
}

Basically, the above code will generate a new file with the desired replace operation, but as you can see, the way I am doing is read each line of the original file and write to a new file. This could achieve my goal, but it may have system IO issue because it is reading and writing for each line. Also, I cannot read all the lines to an array first, and then write, because the file is large and if I try to write to an string[] , replace all, then write the array to the file, will bring about the memory timeout issue.

Is there any way that I can just locate to the specific lines, and just replace those lines and keep all the rest? Or What is the best way to solve the above problem? Thanks

Answer 1

In general, what you are doing is correct, possibly followed by some renames to replace the original file. If you do want to replace the original file, you should rename the original file to a temporary name, rename the new file to the original name, and then either leave or delete the original file. You must handle conflicts with your temporary name and errors in all renames.

Consider you are replacing a six character string with a five character string - if you write back to the original file, what will you do with the extra characters? Files are stored on disk as raw bytes of data, there is no "text" file on disk. What if you replace a string with a longer one - you then potentially have to move the entire rest of the file to make room to write the longer line.

You can imagine the file on disk as letters written on graph paper in the boxes. The end of each line is noted by a special character (or characters - in Windows, that is CRLF), the characters fill all the boxes horizontally. If you tried to replace words on the graph paper you would have to erase and re-write lots of letters. Writing on a new sheet will be easiest.

Answer 2

Well, your approach is basically fine... but I wouldn't check if the line contains the word before ... the trade-off is not good enough:

using (StreamReader sr = new StreamReader(@"S:\Personal Folders\A\TESTA.txt"))
{
    using (StreamWriter sw = new StreamWriter(@"S:\Personal Folders\A\TESTB.txt"))
    {
        String line;

        while ((line = sr.ReadLine()) != null)
            sw.WriteLine(line.Replace("before", "after"));
    }
}

Answer 3

I don't know what IO issue you are worried about, but your code should work ok. You can code more concisely as follows:

using (StreamReader sr = new StreamReader(@"S:\Personal Folders\A\TESTA.txt"))
{
    using (StreamWriter sw = new StreamWriter(@"S:\Personal Folders\A\TESTB.txt"))
    {
        while ((string line = sr.ReadLine())!= null)
        {
            sw.WriteLine(line.Replace("before", "after"));
        }
    }
}

This will run a bit faster because it searches for "before" only once per line. By default the StreamWriter buffers your writes and does not flush to the disk each time you call WriteLine, and file IO is asynchronous in the operating system, so don't worry so much about IO.

Answer 4

Try following :

else if (line.Contains("before"))
{
    sw.WriteLine(line.Replace("before", "after"));
    sw.Write(sr.ReadToEnd());
    break;
}