simple stuff... C# string arrays

C#:
Okay this seems freakishly easy, but all the help files and forums I've seen only show me how to do this with fixed lengths, etc.

I have a large string (~6.5MB) which I want to parse into an array of thousands of smaller strings (=1 line each).  The splitting part seems easy, just seperate them at every instance of "\n".  What I can't figure out is how to declare the array of a dynamic number of strings.  So far what I have is this:

internal static void ArrayCreate(string fileText) {
string[] logEntry;
int[] logIndex, logIndexNext;
for (int tempIndex=1; tempIndex < (fileText.Length); tempIndex++) {
string[] logIndex = new string[tempIndex];
int logIndex[tempIndex]=fileText.IndexOf("\n",tempIndex);
int logIndexNext[tempIndex] = fileText.IndexOf("\n",(logIndex[tempIndex]+1));
}

This seems to me as the way to do it.  What am I doing wrong?  For anyone interested, the project is a reader that will download and parse access.log files from a Yahoo-based webserver, then provide an arseload of information that will be dealt with later.

As will be more obvious to you than to me, I'm getting errors:
1. Syntax error, bad array declarator.
2. Array size cannot be specified in a variable declaration

So what am I doing wrong?

logEntry[] = arrays of strings that will contain an individual log entry
logIndex[] = Pointers for logEntry[]
logIndexNext[] = Pointer for the next entry (logEntry[] not done building yet)
tempIndex = temp pointer for current progress

The string.Split() method should do it for ya...

    string info = "mark\nsmith\n123 csharp drive\ntoronto\ncanada";

    string[] arInfo = info.Split('\n');

Performance and memory usage wise you might be better of reading the log file with a StreamReader using ReadLine instead of reading the entire thing into one huge string and then splitting that. Put the string into a System.Collections.Specialized.StringCollection (or if you're using 2.0 you could use System.Collections.Generic.List<String>) as you read them.

Y'know, Sven, I tried that at first; but it seems that once it read in a few hundred thousand entries it froze up. I found this by adding a simple if-messagebox line. When I switched it to a ReadToEnd() call, it read the entire file in three seconds.  Plus, I intend to have the file grow larger since it's only two weeks' worth of logs.

Beltira1968 wrote:

The string.Split() method should do it for ya...

    string info = "mark\nsmith\n123 csharp drive\ntoronto\ncanada";

    string[] arInfo = info.Split('\n');

That did it, and nice and fast too. Thanks.

SlackmasterK wrote:

C#:

...

What I can't figure out is how to declare the array of a dynamic number of strings.

FYI (and after you've gotten your answer) you can't--arrays are fixed length.

Lists are the closest variable length equivalent to an Array, where you can get elements by index and such.  There is no appropriate language feature in C#, you're stuck using the class library.

With .NET 2.0, you can achieve the same thing just with a single line of code:
String[] logs = System.IO.File.ReadAllLines("logs.txt");

PS: this implemention is quite simple, but it isn't performant just as Sven broken out above.

Sheva

amotif wrote:

FYI (and after you've gotten your answer) you can't--arrays are fixed length.

Strange, I was able to do it just fine:

string[] logEntry = Log.processLog(fileName).Split('\n');

// Log.processLog:

public static string processLog(string fileName) {
System.IO.StreamReader s = System.IO.File.OpenText(fileName);
string fileText = s.ReadToEnd(); // Parse local file into string
s.Close();
return fileText;}

The result (currently) is an array of about 28,000 strings. 

SlackmasterK wrote:

amotif wrote: FYI (and after you've gotten your answer) you can't--arrays are fixed length.

Strange, I was able to do it just fine:

string[] logEntry = Log.processLog(fileName).Split('\n');

Split creates a new array that just so happens to be the right size. You can't add more to it. It's funky like that.

[dupe]

amotif wrote:

SlackmasterK wrote:C#: ... What I can't figure out is how to declare the array of a dynamic number of strings.

FYI (and after you've gotten your answer) you can't--arrays are fixed length.

I suppose I could deconstruct your question a little differently.

a) Yes, you can declare a reference to an array of indescriminant length:

string[] a;
a = new string[0];
a = new string[1];
a = new string[1024*1024];

After all, C# is not Ada.

b) However, as pointed about above arrays are still fixed length.

Carry on.

As an off topic, The File.RealAllLines method is internally defined as follows:

public static string[] ReadAllLines(string path, Encoding encoding)
{
      ArrayList list1 = new ArrayList();
      using (StreamReader reader1 = new StreamReader(path, encoding))
      {
            string text1;
            while ((text1 = reader1.ReadLine()) != null)
            {
                  list1.Add(text1);
            }
      }
      return (string[]) list1.ToArray(typeof(string));
}

Why not use List<String> instead of ArrayList, so that the cast operation can be crossed out in the return statement?

Sheva

Interesting.. This thread has proven most helpful thus far; ReadAllLines has multiplied the log parsing method by a factor of a few hundred.  Not much for status updates in the process, but it's so fast they're not really necessary.

SlackmasterK wrote:

So what am I doing wrong?

A question, am I not correct? In fact, a question that has been asked. Then why:

SlackmasterK wrote:

If you have to ask, you're not ready to know yet.

SlackmasterK wrote:

Y'know, Sven, I tried that at first; but it seems that once it read in a few hundred thousand entries it froze up. I found this by adding a simple if-messagebox line. When I switched it to a ReadToEnd() call, it read the entire file in three seconds.  Plus, I intend to have the file grow larger since it's only two weeks' worth of logs.

SlackmasterK wrote:

Interesting.. This thread has proven most helpful thus far; ReadAllLines has multiplied the log parsing method by a factor of a few hundred.  Not much for status updates in the process, but it's so fast they're not really necessary.

FYI, ReadAllLines does exactly what I described earlier: it goes over the file line by line with a StreamReader.

In fact, according to Reflector this is the code for ReadAllLines:

public static string[] ReadAllLines(string path, Encoding encoding)
{
      ArrayList list1 = new ArrayList();
      using (StreamReader reader1 = new StreamReader(path, encoding))
      {
            string text1;
            while ((text1 = reader1.ReadLine()) != null)
            {
                  list1.Add(text1);
            }
      }
      return (string[]) list1.ToArray(typeof(string));
}

Which is precisely what I described. I probably would've used a List<string> instead of an ArrayList, but that hardly matters in this case.

EDIT: Sorry, didn't see footballism's post.

Khamul wrote:

SlackmasterK wrote: So what am I doing wrong?

A question, am I not correct? In fact, a question that has been asked. Then why:

SlackmasterK wrote: If you have to ask, you're not ready to know yet.

I rotate my captions sometimes.  Unfortunately, the one I really want won't fit:

There is nothing more fearful than ignorance in action. 
Think left and think right, think low and think high;
Oh the things you can think of, if only you try.

footballism wrote:

As an off topic, The File.RealAllLines method is internally defined as follows:

[snip]

Why not use List<String> instead of ArrayList, so that the cast operation can be crossed out in the return statement?

These are just guesses:

• There's no need to change the implementation of BCL methods just because there's another way to do it.
• Perhaps the performance or size implications aren't desirable.