Generics, IEnumerable and LINQ

Generics

• When using lists, you have to define the type of data that will be stored in the list, inside the angled brackets:

  List<int> numberList = new List<int>();     // This list stores variables of type int
  

• This means that list is a generic class: it can contain data of any type
• Classes, structs, interfaces and methods can also be generic
• All the variables that are defined generic inside a generic container will be assigned a type only when the containing object/method is called

Creating a generic class

class GenericClassExample<T>
{
  public T value;    // This value will be whatever type is specified at instantiation
  public void PrintTypeAndValue()
  {
    Console.WriteLine
      ($"This class contains a variable of type {value.GetType()} and of value {value}");
  }
}
class Program
{
  static void Main(string[] args)
  {
    GenericClassExample<int> example = new GenericClassExample<int>();
    example.value = 20;
    example.PrintTypeAndValue();
  }
}

Multiple type parameters

Generic classes can receive multiple types as parameters:

class CustomContainer<T1, T2, T3>
{
  public T1 First { get; set; }
  public T2 Second { get; set; }
  public T3 Third { get; set; }
}
class Program
{
  static void Main(string[] args)
  {
    CustomContainer<int, string, DateTime> container
      = new CustomContainer<int, string, DateTime>();
    container.First = 10;
    container.Second = "Testing.";
    container.Third = DateTime.Now;
  }
}

Creating a generic Method

void GenericMethodExample<T>(T value)
{
  Console.WriteLine
    ($"This method was passed a variable of type {value.GetType()} and of value {value}.");
}
GenericMethodExample<string>("ABC");

• Note: You could name the generic type as anything, e.g. <GenericType>. It is named <T> by convention.

Exercise 1: Initializing a populated list

Create a generic method GetPopulatedList<T> which takes two parameters: T value and int length, and returns a new list of type T which is populated with the value variables and has a length of length.

Test your method out with a couple of different types and lengths:

List<string> list = GetPopulatedList<string>("Hello, there", 10);
foreach(string value in list)
{
    Console.WriteLine(value);
}

IEnumerable

• Lists and arrays are both collections that implement the IEnumerable interface
• All objects that implement the IEnumerable interface can be iterated with the foreach statement

  IEnumerable<string> names = new string[] {"Harry", "Luke", "Harley"};
  IEnumerable<string> days = new List<string> {"Sunday", "Monday", "Friday"};
  
  foreach (string name in names)
    Console.WriteLine(name);
  
  foreach (string day in days)
    Console.WriteLine(day);
  

• The IEnumerable interface itself doesn't hold much functionality:
  
• However, the LINQ library includes all the methods you would typically need to apply to IEnumerables, such as filtering

LINQ

The problem

• Here's an example of a common query a programmer might have to do:

• Try to find an object with a certain id from an array

  IEnumerable<User> userArray = new User[2]     // Initialize a new array of users
  {
    new User { Id = 0, Name = "Rene" } ,
    new User { Id = 1, Name = "Ville" }
  };
  
  User FindObjectWithId(int id)
  {
    foreach (User user in userArray)
    if (user.Id == id)
      return user;
    return null;
  }
  Console.WriteLine(FindObjectWithId(1).Name);    // Outputs "Ville"
  

• Having to write your own method for every possible query operation (select, filter, sort…) would be nonsensical
• The List class includes some methods for manipulation, but...
  • These only work on lists; not all IEnumerables (e.g. arrays) contain those methods!
• Not to mention different data types altogether (objects, SQL databases, XML, JSON…)
• What to do?

The solution

• To introduce extensive query capabilities to all collection types, Language-Integrated Query (LINQ) was created
• LINQ supports querying of objects and even XML and SQL data, directly in your code
• Get started by adding the System.Linq namespace to your project:

  using System.Linq;
  

• Here's a solution to the earlier example that uses LINQ:

  IEnumerable<User> userArray = new User[2]      // Initialize a new array of users
  {
    new User { Id = 0, Name = "Rene" } ,
    new User { Id = 1, Name = "Ville"}
  };
  Console.WriteLine(userArray.First(user => user.Id == 1).Name);    // Outputs "Ville"
  

• The Enumerable.First method returns the first result that satisfies the expression in the parameters

• The arrow syntax above is called a lambda expression

Lambda expressions

• Lambda expressions are a quick way of writing one-line methods
• The First method of LINQ takes a delegate as a parameter
• We could just declare a method beforehand, and then refer to it in First like this:

  bool GetUserWithId(User user)
  {
    return user.Id == 1;
  }
  Console.WriteLine(userList.First(GetUserWithId).Name);    // Outputs "Ville"
  

• To shorten the expression, we could use a delegate inside the parameters to refer to an anonymous method:

  Console.WriteLine(userList.First(delegate (User user)
    { return user.Id == 1; }));    // Outputs "Ville"
  

Lambda expressions compared

Using LINQs "First" -query with...

bool GetUserWithId(User user)
{
  return user.Id == 1;
}
Console.WriteLine(
  userList.First(
    GetUserWithId).Name);
// Outputs "Ville"

Console.WriteLine(userList.First(
delegate (User user)
{
  return user.Id == 1;
}));
// Outputs "Ville"

Console.WriteLine(userList.First(
  user =>
    user.Id == 1).Name);
// Outputs "Ville"

Lambda expressions: An example

Regular methods can also be declared using the arrow function

static void Main(string[] args)
{
  // Method body assigned with lambda expression
  string PrintCheckUpper(bool upper, string text) =>
    upper ? text.ToUpper() : text;
  
  PrintCheckUpper(true,  "I'm not angry!");    // Outputs I'M NOT ANGRY!
  PrintCheckUpper(false, "I'm not angry!");    // Outputs I'm not angry!
}

LINQ and Lambda expressions

• Going back to our LINQ example...

  IEnumerable<User> userArray = new User[2]      // Initialize a new array of users
  {
    new User { Id = 0, Name = "Rene" } ,
    new User { Id = 1, Name = "Ville"}
  };
  

• Here's a solution without lambda expressions:

  bool GetUserWithId(User user)
  {
    return user.Id == 1;
  }
  Console.WriteLine(userArray.First(GetUserWithId).Name);    // Outputs "Ville"
  

• And here's the shorter version with lambda expression:

  Console.WriteLine(userArray.First(user => user.Id == 1).Name);    // Outputs "Ville"
  

LINQ methods

• LINQ contains methods for filtering, ordering, grouping, joining and selecting
• Suppose we have a class Person that contains a property Country
• The following performs filtering to the persons object, returning only the persons whose country is of value "Finland":

  List<Person> persons = new List<Person> {/* Insert data here */};
  
  var queryResult = persons
    .Where(person => person.Country == "Finland");
  

LINQ methods listed

More LINQ methods can be found here

LINQ: A chained example

• Suppose we have a csv file authors.csv in our project directory.

  // using System.IO;
  string path = @"C:\some\path\authors.csv";
  var authors = File.ReadAllLines(path)
    .Skip(1)         // Skip the first line which contains the column titles
    .Select(line =>  // Using lambda expression, return the new Author objects
    {
      var columns = line.Split(',');
      return new Author
      {
        Id = int.Parse(columns[0]),
        Name = columns[1],
        Description = columns[2]
      };
    })
    .ToList();       // Turn the resulting IEnumerable into a list
  

• authors now contains all the authors from the original csv file!

Query syntax

• An alternative way of using LINQ is with the query syntax
• The following performs filtering to a persons object:

  var queryResult = from person in persons
    where person.country == "Finland"
    select person;
  

• The range variable person works like the range variable in a foreach loop in that it holds a copy of the current element in the persons variable
• The where keyword specifies the condition for the filter

LINQ queries: An example

• Suppose we have a class City that contains a property Description
• The following would print the descriptions of all cities that have descriptions less than 100 characters long

  IEnumerable Cities = new List<City> {/* Insert data here */};
  
  var filteredResult = from city in Cities
    where city.Description.Length < 100
    select city;
  
  foreach (City result in filteredResult)
    Console.WriteLine(result.Description);
  

Extension methods vs. query syntax

• In the previous examples, we used the query syntax of LINQ
• Both the methods and queries do pretty much the same thing
• It is up to you which syntax you want to use
  • The method syntax works like any normal C## methods
  • The query syntax might be more approachable to those who are familiar with SQL

var methodResult = persons
  .Where(person => 
    person.Country == "Finland");

var queryResult = from person in persons
  where person.Country == "Finland"
  select person;

• All queries listed here: https://www.tutorialsteacher.com/linq/linq-standard-query-operators

ToArray() and ToList() methods

• Notice that the LINQ queries return an IEnumerable
• If you need to use arrays or lists, you need to call the ToArray() or ToList() methods

string[] strings = new string[] 
  { "Timo", "Pekka", "Taina", "Kalle" };
string[] queryResult = strings
  .Where(s => s.StartsWith('T'));

string[] strings = new string[]
  { "Timo", "Pekka", "Taina", "Kalle" };
string[] queryResult = strings
  .Where(s => s.StartsWith('T'))
  .ToArray();

Exercise 2: Filtering Names

• Download this file of names and add it to your project folder: https://raw.githubusercontent.com/dominictarr/random-name/master/names.txt
• Read all the contents into a string array with File.ReadAllLines()
• Create a main loop where the user is asked for a string. Print the total number of names which contain that string.
• If there are less than 10 resulting names, print the names as well!

Exercise 3: Queries on Object Lists

• Expand on the exercise 2.
• Create a new class User with two properties, int Id and string Name
• If the number of filtered names is less than 10, create a list of Users with those names and a running Id
• Sort the list of users by the length of the Name property
• Print the names and id:s of the users in the sorted list

Going Further: Extension Methods

• Recap: IEnumerable itself only contains one method
• How does the LINQ library suddenly add all these methods to our Enumerables?
• This is possible with extension methods!