Using Deconstructors in C# 7.0

This is another awesome feature that got introduced with C# 7.0 which can be helpful in many scenarios while consuming various objects. Deconstruction is mostly discussed with another new feature System.ValueTuple but it can be used with other custom type as well. I have also discussed it in briefly in ValueTuple post at the link below. I will advise to go through that first before starting here.

How to use the new ValueTuples : A C# 7.0 feature

What is Deconstruction

It is a process of splitting an instance’s (value types and reference types) value into multiple parts and assigning them to new variables. Like, I mentioned that it is extremely useful with the Tuples which contains a set of values but can be used with other types.

How to use Deconstruction

We can leverage it in our custom classes, structs or interfaces. We need to implement Deconstruct method, which returns void and each deconstructed value should be added as out parameter in the method signature. We can have multiple overridden Deconstruct method based on the requirement. Let’s see an example

public class Person
{
    public string FirstName { get; set; }

    public string MiddleName { get; set; }

    public string LastName { get; set; }

    public int Age { get; set; }

    public Person(string fName, string mName, string lName, int age)
    {
        FirstName = fName;
        MiddleName = mName;
        LastName = lName;
        Age = age;
    }
    public void Deconstruct(out string fName, out string lName)
    {
        fName = FirstName;
        lName = LastName;
    }

    public void Deconstruct(out string fName, out string lName, out int age)
    {
        fName = FirstName;
        lName = LastName;
        age = Age;
    }
}

Here we can see that I have provided two Deconstruct methods, one with two out parameters: fName, lName and other with three : fName, lName, age.

Person p = new Person("Brij", "Bhushan", "Mishra", 32);

//(string firstName, string lastName) = p;
// OR
var (firstName, lastName) = p;

Console.WriteLine($"Person Details: First Name: {firstName}, Last Name: {lastName}");


(string firstName, string lastName, int age) = p;

Console.WriteLine($"Person Details: First Name: {firstName}, Last Name: {lastName}, Age: {age}");

Here I created the instance of Person with all the details. Then in first deconstruction, we got the first name and last name which calls the first Deconstruct method of the class which has two out parameters. Note that we have two ways of deconstruction. First we explicitly declare the type of each field inside parentheses (commented out) and another, using var keyword so that C# infers the type of each variable. Similarly later, second Deconstruct method is called which deconstructs it in three variables. It can be really useful when we have a class with numerous properties, fields and most of the time, we need only few there we can use the Deconstruct methods.

Similar to ValueTuples, we can use discards here as well. To discard a variable of Deconstruct method, we can use underscore (_) as

Person p = new Person("Brij", "Bhushan", "Mishra", 32);
(string firstName, _ , int age) = p;

Console.WriteLine($"Hello {firstName}, your age is {age}");

Deconstruction use Extension method

There could be scenarios where you want to this feature but its an existing class where changes are not possible (like some third party classes, classes own by different team or even framework classes), there we can add this as an extension method. Lets assume that Person class doesn’t have Deconstruct methods and we can add it as an extension method as

public static class DeconstructionExtensions
{
    public static void Deconstruct( this Person p, out string fName, out string lName, out int age)
    {
        fName = p.FirstName;
        lName = p.LastName;
        age = p.Age;
    }
}

It will work in the same way as earlier.

Hope you have enjoyed the post and will be able to use in your day to day coding.

Thanks
Brij

C# 6.0 and 7.x posts links

Hello All,

There are rapid changes to C# language recently. Although these are very useful in writing cleaner, concise, less repetitive and performance oriented code, I see very less usage of these in our day to day coding. So I thought of writing these features as blog posts to learn myself and share with others. In this post I am just listing down the links of my earlier C#6.0 and C# 7.x blog posts so that it is easily available. Do let me know if there is any important feature that I missed which could be useful in our day to day coding.

C# 6.0 features

C# 7.X features

Hope you’ll enjoy these posts and able to use these in your day to day development.

Cheers,
Brij

TransactionScope: A simple way to handle transactions in .NET

Have you ever tried implementing transactions using C# code? Normally, we implement transactions in SQL where multiple Insert/Update statements takes part in it. A Transaction follows the ACID (Atomicity, Consistency, Isolation, Durability) rule where either all the statements get committed or all get canceled and rolled back. TransactionScope allows us to implement it at application level. There could be some scenarios where you are required to do different operations in the same database or even multiple databases (distributed transaction) or due to some other constraints, it cannot be done at database level. It is also very helpful for application developers if they have less exposure to database.

What is TransactionScope

TransactionScope got introduced with .NET 2.0 as part of  System.Transaction. It is a class which provides a simple way to make a set of operations as part of a transaction without worrying about the complexity behind the scene. If any of the operation fails in between, entire transaction would fail and rolled back  which undo all the operation that got completed. All this would be taken care by the framework, ensuring the data consistency.

How to use the TransactionScope?

To use this, you need to add the reference of System.Transactions reference which is part of framework libraries (normally it wont be added by default). Once it get added, add the namespace System.Transactions wherever we want to use this. The syntax would look as

try
{
    using (TransactionScope scope = new TransactionScope())
    {
        // Do Operation 1
        // Do Operation 2
        //...

        // if all the coperations complete successfully, this would be called and commit the trabsaction. 
        // In case of an exception, it wont be called and transaction is rolled back
        scope.Complete();
    }
}
catch (ThreadAbortException ex)
{
    // Handle exception
}

Here we can see that we have used Disposable block while creating instance of TransactionScope, it makes sure the dispose gets called when it gets out of the block and ends the transaction scope.

In one Transaction scope, we can do multiple operation connecting to different databases as

using (TransactionScope scope = new TransactionScope())
{
    using (con = new SqlConnection(conString1))
    {
        con.Open();

        // Do Operation 1
        // Do Operation 2
        //...
    }

    using (con = new SqlConnection(conString2))
    {
        con.Open();

        // Do Operation 1
        // Do Operation 2
        //...

    }

    scope.Complete();
}

Here we are using two connection strings to connection different databases. We can use as many based on our requirement. We can have nested transactions as well. It could be as

public void DoMultipleTransaction()
{       
    try
    {
        using (TransactionScope scope = new TransactionScope())
        {
            using (con = new SqlConnection(conString1))
            {
                con.Open();
                // Do Operation 1
            }

            OtherTransaction();
            scope.Complete();
        }
    }
    catch (ThreadAbortException ex)
    {
        // Handle exception
    }
}

private void OtherTransaction()
{
    using (TransactionScope scope = new TransactionScope())
    {
        using (con = new SqlConnection(conString2))
        {
            con.Open();
            // Do Operations
        }
        scope.Complete();
    }
}

Here the outermost transaction is called as rootscope and here even if the inner transaction (OtherTransaction above) gets completed by calling scope.Complete(), if the rootscope complete could not be called due to various reasons, then the complete transaction would be rolled back including inner transactions.

Note: You might get one of the following exception while executing distributed trsanctions

  1. MSDTC on server is unavailable
  2. Network access for Distributed Transaction Manager (MSDTC) has been disabled.

 

Both the error are due to the same reason, first one occurs when you have the database and the application the same server while 2 if on the other server. For same server, go to run-> cmd-> services.msc. Run the service named Distributed Transaction Coordinator and make the startup type automatic so that it gets started again in case of system restart. For 2, follow the link to cofigure MSDTC.

TransactionScope provides various TransactionScopeOptions which defines transactions behavior for the scope. Lets see an example

using (TransactionScope scope = new TransactionScope())
{
    // Do Operation
    using (TransactionScope scope1 = new TransactionScope(TransactionScopeOption.Required))
    {
        // Do Operation
        scope1.Complete();
    }
    using (TransactionScope scope2 = new TransactionScope(TransactionScopeOption.RequiresNew))
    {
        // Do Operation
        scope2.Complete();
    }
    using (TransactionScope scope3 = new TransactionScope(TransactionScopeOption.Suppress))
    {
        // Do Operation
        scope3.Complete();
    }

    scope.Complete();
}

Here we created three transactions under the parent transaction with different TransactionScopeOptions. By default the scope is required, which applies to parent transaction here. It is a rootscope which creates a new transaction, and mark it as an ambient transaction. scope1 is also created with required and as we have already an ambient transaction (scope) so it joins the parent transaction. scope2 got created with option as RequiresNew which means it is a new transaction which is independently works with ambient transaction. scope3 got created with suppress option, which means it doesn’t take part in any ambient transactions. It gets executed regardless whether ambient transaction executes successfully or not. All the ambient transactions gets committed once the parent (global) scope completes.

Hope you enjoyed this post and will be using transaction in your future requirements.

Cheers,
Brij

Expression-bodied Members in C#

Expression-bodied members are one of the shiny features of C# 6.0 which allows us to write the implementation of a member in more readable and concise format. We can use the expressions as a definition of the members instead of using the statement block. The format of the expression is as

var=> expression

This is also called lambda expressions and => is called lambda operator. In C# 6.0, we can use this feature with methods and properties. Lets see few examples

Properties: Instead of using the normal getter, we can use expressions with lambda operators as mentioned.


// Normal way
public string FullName
{
    get { return string.Format("{0} {1}", FirstName, LastName); }
}

// Using Expression (C# 6.0)
public string FullName => string.Format("{0} {1}", FirstName, LastName);

// Above code can be more consize using string interpolation
public string FullName =>$"{FirstName} {LastName}";

If you are not aware of String Interpolation, you can refer one of my previous post here.

Methods : Similar to properties, C# 6.0 also allows us to use expressions while writing methods which returns values to the caller.


// Normal way
public string GetFullName(string firstname, string middleName, string lastName)
{
    return middleName == null ? $"{firstname} {lastName}" : $"{firstname} {middleName} {lastName}";
}

// Using Expessions 
public string GetFullName(string firstname, string middleName, string lastName) => middleName == null ? 
             $"{firstname} {lastName}" : $"{firstname} {middleName} {lastName}";

Limitation: There is a limitation of expression bodied members.We can have just single statement in the expression and statement blocks {} are not allowed. In my above example, I used the ternary operator conditional logic instead of using if statement block.

As we can see C# 6.0 added the capability for methods and read only properties, C# 7.0 added more power to this feature and now we can use expressions in constructors, set accessors in properties, indexers and finalizers. Let’s take a look

Properties: We saw earlier enhancements for Property which was similar to get accessor

private string _name;
public string Name {
    get => _name;
    set => _name = value;
}

Here this is a very simple example and in this scenario, I would like to prefer the Auto propery as below

public string Name { get; set; }

However in my earlier example where I was returning FullName after concatenating two strings, there it is better suited. So when you are doing some smaller operations as well in your property then it is better choice.

Constructor: We can leverage the expression with the constructors as well. Let’s see

// Earlier
public Person(string name)
{
    this.Name = name;
}

// With C# 7.0
public Person(string name) => this.Name = name; 

As mentioned earlier that statement block can’t be used here. So say if I have multiple parameters to pass so I can use another C# 7.0 features: ValueTuple and Destructors.

public Person(string firstName, string lastName) => (FirstName, LastName) = (firstName, lastName);

I wrote some time back on ValueTuples and Deconstructors, you can take a look.

Finalizers: We can use expressions in finalizers as well. Let’s see

// Earlier
~Person()
{
    Console.WriteLine("Person's destructor");
}

// Using Expressions
~Person() => Console.WriteLine("Person's destructor");

There are many new features that got intrudocuced in C# 6.0 and C# 7.X (7.0, 7.1, 7.2) if these are used together, can provide lot of value. The whole idea is write cleaner, concise and more performant code so that the focus is more on business logic.

Hope you are enjoying the enhancements.

Cheers
Brij