Question

I would like to modify the following code to work with large files.

    public static void Replace(string filePath, string searchText, string replaceText)
    {
        StreamReader reader = new StreamReader(filePath);
        string content = reader.ReadToEnd();
        reader.Close();

        content = Regex.Replace(content, searchText, replaceText);

        StreamWriter writer = new StreamWriter(filePath);
        writer.Write(content);
        writer.Close();
    }

I’m thinking that I’m going to need to open a filestream to write to a new file name, and then delete the original file and replace it with the new one when I’m done. Does that sound about right?

Also…
I love the simplicity of this routine, that aside from the necessary file i/o lines of code there is only one line of code to process the file. However, I’m also wondering if I’m sacrificing performance for simplicity… is Regex.Replace very performant?

Answer 1

First: you can try Regex with Stream (it seems be more faster and less memory requirement):

• Building a Regular Expression Stream Search with the .NET Framework

or see Mono-Project Regex. It has Regex with streaming.

see this article for Regex performance:

• C# Regex Performance (from dotnetperls)

or if using Regex is not necessary use String.Replace and try this one line code:

File.WriteAllText(filePath, 
                  File.ReadAllText(filePath).Replace(searchText, replaceText));