Question

I have a text file exported from a Foxpro (Dos-based) program, but this text contains non-English characters ( arabic [ right to left ] ) and now exported strings are like this “¤“îگüَن” .

is there any way to convert them back to their original values ?

Answer 1

You should read the data with the proper code page.

public static string ReadFile(string path, int codepage)
{
    return Encoding.GetEncoding(codepage)
        .GetString(File.ReadAllBytes(path));
}

Invoke the function with the right code page ID, for MS-DOS Arabic it should be “708”, for a complete list you can start on Wikipedia.

string content = ReadFile(@"c:\test.txt", 708);

Solution with look-up table to translate from unsupported encodings (mapping is required only for charactes > 127):

public static string ReadFile(string path, byte[] translationTable, int codepage)
{
    byte[] content = File.ReadAllBytes(path);
    for (int i=0; i < content.Length; ++i)
    {
        byte value = content[i];
        if (value > 127)
            content[i] = translationTable[value - 128];
    }

    return Encoding.GetEncoding(codepage)
        .GetString(content);
}

An example of translation table:

Index    Original (IS)    Translated (1256)
...
13       141              194
...