Basically, don't try and work with the file as lines of text; separate that away, so that you have one piece of code which parses that text into typed records, and then process those upstream when you only need to deal with typed data.
For example (and here I'm assuming that the file is tab-delimited, but you could change it to be column indexed instead easily enough); look at how little work my Main method needs to do to work with the data:
using System;
using System.Collections.Generic;
using System.Globalization;
using System.IO;
using System.Linq;
static class Program
{
static void Main()
{
foreach (var item in ReadFile("my.txt").OrderBy(x => x.Joined))
{
Console.WriteLine(item.Names);
}
}
static readonly char[] tab = { '\t' };
class Foo
{
public string Names { get; set; }
public int Age { get; set; }
public string Designation { get; set; }
public DateTime Joined { get; set; }
}
static IEnumerable<Foo> ReadFile(string path)
{
using (var reader = File.OpenText(path))
{
// skip the first line (headers), or exit
if (reader.ReadLine() == null) yield break;
// read each line
string line;
var culture = CultureInfo.InvariantCulture;
while ((line = reader.ReadLine()) != null)
{
var parts = line.Split(tab);
yield return new Foo
{
Names = parts[0],
Age = int.Parse(parts[1], culture),
Designation = parts[2],
Joined = DateTime.Parse(parts[3], culture)
};
}
}
}
}
And here's a version (not quite as elegant, but working) that works on .NET 2.0 (and probably on .NET 1.1) using only ISO-1 language features; personally I think it would be silly to use .NET 1.1, and if you are using .NET 2.0, then List<T> would be vastly preferable to ArrayList. But this is "worst case":
using System;
using System.Collections;
using System.Globalization;
using System.IO;
class Program
{
static void Main()
{
ArrayList items = ReadFile("my.txt");
items.Sort(FooByDateComparer.Default);
foreach (Foo item in items)
{
Console.WriteLine(item.Names);
}
}
class FooByDateComparer : IComparer
{
public static readonly FooByDateComparer Default
= new FooByDateComparer();
private FooByDateComparer() { }
public int Compare(object x, object y)
{
return ((Foo)x).Joined.CompareTo(((Foo)y).Joined);
}
}
static readonly char[] tab = { '\t' };
class Foo
{
private string names, designation;
private int age;
private DateTime joined;
public string Names { get { return names; } set { names = value; } }
public int Age { get { return age; } set { age = value; } }
public string Designation { get { return designation; } set { designation = value; } }
public DateTime Joined { get { return joined; } set { joined = value; } }
}
static ArrayList ReadFile(string path)
{
ArrayList items = new ArrayList();
using (StreamReader reader = File.OpenText(path))
{
// skip the first line (headers), or exit
if (reader.ReadLine() == null) return items;
// read each line
string line;
CultureInfo culture = CultureInfo.InvariantCulture;
while ((line = reader.ReadLine()) != null)
{
string[] parts = line.Split(tab);
Foo foo = new Foo();
foo.Names = parts[0];
foo.Age = int.Parse(parts[1], culture);
foo.Designation = parts[2];
foo.Joined = DateTime.Parse(parts[3], culture);
items.Add(foo);
}
}
return items;
}
}
File.ReadAllLinesinstead which returns aString[]? Even better, load aList<CustomClass>whereCustomClasshas the properties above.foreach(var item in items) Console.WriteLine(item);myEnumeratorisIDisposable, and if so ensure you dispose it (for both success and failure) -foreachdoes a lot of good things for you. Additionally,foreachsupports duck-typing, meaning: it can be more efficient thanIEnumerator