reading and editing contents of file

Question

I have a file which is in this format:

AA
AA AA AA
AA AA AA AA AA AA
AA AA AA
file.txt
0

Where AA represents a one or two digit number (the length of each doesn't have to be 2), and file.txt is the name of a file. I need a very fast method to obtain the file name. Also, I need to replace the final 0 with 1. If there is no 0 in the file at the end (its optional), I need to append 1 on a new line. Here is my current code:

StringBuilder sb = new StringBuilder("\r\n1");
string txt = File.ReadAllText(args[0]);
string[] lines = txt.Split('\n');
string name = lines[4];

if (lines.Length != 6) // Check if EOF is 0 or not.
    txt += sb.ToString();
else
    txt = txt.Substring(0, txt.Length - 1) + '1'; // Replace 0 with 1.

File.WriteAllText(args[0], txt);
Console.WriteLine(name);

However, I was wondering if there was even a faster way to do this (without having to load the file in memory maybe). I have to process tons of files like this, so saving tenths would be extremely useful.

Are you getting the filename from the text inside the file or do you want it from its name on the os? — Bali C, Dec 23 '11 at 15:13
@BryceAtNetwork23: Yeah, and if there are 6 lines then I'm certain there will be 0 or 1 at the end. — dnclem, Dec 23 '11 at 15:16
@BaliC: I'm getting the filename from inside this file, not from anywhere else. — dnclem, Dec 23 '11 at 15:16
Why are you using a StringBuilder just to hold const? It's completely unnecessary and will slow you down. Just do `txt += "\r\n1";` — Igby Largeman, Dec 23 '11 at 15:25
My guess is that if you are dealing with less than hundred of thousands of file, performance will not be an issue — Eric, Dec 23 '11 at 15:25
@Charles: You're right. I have no clue why I didn't do that in the first place. — dnclem, Dec 23 '11 at 17:30
@Eric: But still, it would be to nice to learn how to optimize code. — dnclem, Dec 23 '11 at 17:30
I wonder which person downvoted this question; at the very least make a comment explaining why. — dnclem, Dec 31 '11 at 05:52

Eugen Rieck · Answer 1 · 2011-12-23T15:25:53.080

1

Loading the complete file into memory will most likely be your fastest choice if filesize is below min(MemoryPageSize, FSBlockSize), typically 4K

So assuming You have the contents of the file in a string, methinks using something like

int n;
if (content.EndsWith("\r\n0")
{
  n=content.Length-3;
  content=content.Substring(0,n+2)+"1";
}
else
{
  n=content.Length;
  content=content+"\r\n1";
}

and write the String out to the file. This will save you from the rather expensive Split()

for the filename we continue as follows:

int p=content.LastIndexOf('\n',n);
String filename=content.Substring(p+1,n-p);

edited Dec 23 '11 at 15:25

answered Dec 23 '11 at 15:20

Eugen Rieck

64,175
10
70
92

Works pretty good. I did not know Split() would slow things down a bit. – dnclem Dec 23 '11 at 17:29

Ivo · Accepted Answer · 2011-12-23T16:39:20.230

1

var lines = new List<string>(File.ReadAllLines(args[0]));
string name = lines[4];

if (lines.Length != 6) // Check if EOF is 0 or not.
    lines.Add("1");
else
    lines[5] = "1";

File.WriteAllLines(args[0], lines);
Console.WriteLine(name);

edited Dec 23 '11 at 16:39

answered Dec 23 '11 at 15:32

Ivo

8,172
5
27
42

There is no method called Add in string[]. – dnclem Dec 23 '11 at 15:40
But there is no ToList method in NET 2.0 and that's what I'm using. :s – dnclem Dec 23 '11 at 16:00
Oh. Of course. I completely forgot about List. Thanks, it works great now. :) – dnclem Dec 23 '11 at 17:29

score 1 · Answer 3 · answered Dec 23 '11 at 15:32

1

If you're looking for speed, then you want to use FileStream.Read to a temporary buffer. Here is something I threw together that might help point you in a slightly different but faster direction:

StringBuilder filename = new StringBuilder();
using (FileStream stream = new FileStream(args[0], FileMode.Open, FileAccess.Read))
{
    byte[] buffer = new byte[1];
    while (stream.Read(buffer, 0, 1) > 0)
    {
        char c = (char) buffer[0];
        if (char.IsLetter(c) || char.IsPunctuation(c))
            filename.Append(c);
    }
}

answered Dec 23 '11 at 15:32

Nick

5,875
1
27
38

I forgot to mention, the numbers AA are actually in hexadecimal so they can include A-F. If that happens, the filename won't always come correctly. – dnclem Dec 23 '11 at 15:46
If that's the case you would have to do some conversion of hext to string – MethodMan Dec 23 '11 at 19:57

reading and editing contents of file

3 Answers3

Linked