9.3. Reading Files¶

9.3.1. Reading to End of Stream¶

In first_file.cs, we explicitly coded reading two lines. You will often want to process each line in a file, without knowing the total number of lines while programming. This means that files provide us with our second kind of a sequence: this time the sequence of lines in the file! To process all of them will require a loop and a new test to make sure that you have not yet come to the end of the file’s stream: You can use the EndOfStream property. It has the wrong sense (true at the end of the file), so we negate it, testing for !reader.EndOfStream in the example program print_file_lines.cs. This little program reads and prints the contents of a file specified by the user, one line at a time:

using System;
using System.IO;

namespace IntroCS
{
   class PrintFileLines  // demo of using EndOfStream test
   {
      public static void Main()
      {
         string userFileName = UI.PromptLine("Enter name of file to print: ");
         var reader = new StreamReader(userFileName);
         while (!reader.EndOfStream) {
            string line = reader.ReadLine();
            Console.WriteLine(line);
         }
         reader.Close();
      }
   }
}

var

For conciseness (and variety) we declared reader using the more compact syntax with var:

var reader = new StreamReader(userFileName);

You can use var in place of a declared type to shorten your code with a couple of restrictions:

• Use an initializer, from which the type of the variable can be inferred.
• Declare a local variable inside a method body or in a loop heading.
• Declare only a single variable in the statement.

We could have used this syntax long ago, but as the type names become longer, it is more useful!

Things to note about reading from files:

• Reading from a file returns the part read, of course. Never forget the side effect: The location in the file advances past the part just read. The next read does not return the same thing as last time. It returns the next part of the file.
• Our while test conditions so far have been in a sense “backward looking”: We have tested a variable that has already been set. The test with EndOfStream is forward looking: looking at what has not been processed yet. Other than making sure the file is opened, there is no variable that needs to be set before a while loop testing for EndOfStream.
• If you use ReadLine at the end of the file, the special value null (no object) is returned. This is not an error, but if you try to apply any string methods to the null value returned, then you get an error!

Though print_file_lines.cs was a nice simple illustration of a loop reading lines, it was very verbose considering the final effect of the program, just to print the whole file. You can read the entire remaining contents of a file as a single (multiline) string, using the StreamReader method ReadToEnd. In place of the reading and printing loop we could have just had:

string wholeFile = reader.ReadToEnd();
Console.Write(wholeFile);

ReadToEnd does not strip off a newline, like ReadLine does, so we do not want to add an extra newline when writing. We use the Write method instead of WriteLine.

9.3.2. Example: Sum Numbers in File¶

We have summed the numbers from 1 to n. In that case we generated the next number i automatically using i++. We could also read numbers from a file containing one number per line (plus possible white space):

static int CalcSum(string filename)
{
   int sum = 0;
   var reader = new StreamReader(filename);
   while (!reader.EndOfStream) {
      string sVal = reader.ReadLine().Trim();
      sum += int.Parse(sVal);
   }
   reader.Close();
   return sum;
}

Below and in sum_file/sum_file.cs is a more elaborate, complete example, that also exits gracefully if you give a bad file name. If you give a good file name, it skips lines that contain only whitespace.

using System;
using System.IO;

namespace IntroCS
{
   class SumFile  // sum a file integers, one per line
   {
      static void Main()
      {
         string filename = UI.PromptLine(
                              "Enter the name of a file of integers: ");
         if (File.Exists(filename)) {
            Console.WriteLine("The sum is {0}", CalcSum(filename));
         }
         else {
            Console.WriteLine("Bad file name {0}", filename);
         }
      }

      /// Open, read and close the named file and
      /// return the sum of an int from
      /// each line that is not just white space.
      static int CalcSum(string filename)
      {
         int sum = 0;
         var reader = new StreamReader(filename);
         while (!reader.EndOfStream) {
            string sVal = reader.ReadLine().Trim();
            if (sVal.Length > 0) {
               sum += int.Parse(sVal);
            }
         }
         reader.Close();
         return sum;
      }
   }
}

A useful function used in Main for avoiding filename typo errors is File.Exists in the System.IO namespace

bool File.Exists(string filenamePath)

It is true if the named files exists in the operating system’s file structure.

You should see the file sum_file/numbers.txt in the Xamarin Studio project. It is in the right form for the program. If you run the program and enter the response:

numbers.txt

you should be told that the file does not exist. Recall that the executable created by Xamarin Studio is two directories down through bin to Debug. This is the default current directory when Xamarin Studio runs the program. You can refer to a file that is not in the current directory. A full description is in the next section, but briefly, what we need now: The symbol for the parent directory is ... The hierarchy of folders and files are separated by \ in Windows and / on a Mac, so you can test the program successfully if you use the file name: ..\..\numbers.txt in Windows and ../../numbers.txt on a Mac. On a Mac, running the program looks like:

Enter the name of a file of integers: ../../numbers.txt
The sum is 16

In FIO Helper Class we will discuss a more flexible way of finding files to open, that works well in Xamarin Studio and many other situations.

9.3.3. Example Copy to Upper Case¶

Here is a simple fragment from example file copy_upper/copy_upper.cs. It copies a file line by line to a new file in upper case:

         var reader = new StreamReader("text.txt");
         var writer = new StreamWriter("upper_text.txt");
         while (!reader.EndOfStream) {
            string line = reader.ReadLine();
            writer.WriteLine(line.ToUpper());
         }
         reader.Close();
         writer.Close();

You may test this in the Xamarin Studio example project copy_upper:

1. Expand the copy_upper project in the Solution pad. The project includes the input file. You may not see it at first. You need to expand the folder for bin and then Debug. You see text.txt.
2. To see what else is in the file system directory, you can select the Debug folder and right click and select Open Containing Folder. You should see text.txt but not upper_text.txt. Leave that operating system file folder open.
3. Go back to Xamarin Studio and run the project. Now look at the operating system Debug folder again. You should see upper_text.txt. You can open it and see that it holds an upper case version of the contents of text.txt.

This is another case where the ReadToEnd function could have eliminated the loop. [2]

string contents = reader.ReadToEnd();
writer.Write(contents.ToUpper());