There are 23 classic design patterns, which are described in the original book,
Design Patterns: Elements of Reusable Object-Oriented Software. These patterns provide solutions to particular problems, often repeated in the software development.
In this article, I am going to describe the how the Command Pattern; and how and when it should be applied.
Command Pattern: Basic Idea
In object-oriented programming, the command pattern is a behavioral design pattern in which an object is used to encapsulate all information needed to perform an action or trigger an event at a later time. This information includes the method name, the object that owns the method and values for the method parameters — Wikipedia
Encapsulate a request as an object, thereby letting you parameterize clients with different requests, queue or log requests, and support undoable operations — Design Patterns: Elements of Reusable Object-Oriented Software
In this pattern an abstract
Command class is declared as an interface for executing operations. The
Command class defines a method named
execute, which must be implemented in each concrete command. This
execute method is a bridge between a
Receiver object and an action. The
Receiver knows how to perform the operations associated with a request (any class may be a
Receiver). Another relevant component in this pattern is the
Invoker class which asks for the command that must be executed.
The UML diagram for this pattern is the following one:
The Command Pattern should be used when:
- You need a command to have a life span independent of the original request. Furthermore, if you want to queue, specify and execute requests at different times.
- You need undo/redo operations. The command’s execution can be stored for reversing its effects. It is important that the
Commandclass implements the methods undo and redo.
- You need to structure a system around high-level operations built on primitive operations.
The Command Pattern has several advantages, summarised in the following points:
- It decouples the classes that invoke the operation from the object that knows how to execute the operation
- It allows you to create a sequence of commands by providing a queue system
- Implementing extensions to add a new command is easy and can be done without changing the existing code.
- You can also define a rollback system with the Command pattern, like in the Wizard example, we could write a rollback method.
- Have strict control over how and when commands are invoked.
- The code is easier to use, understand and test since the commands simplify the code.
Agent which defines the attributes:
stockTrade; and an operation
placeOrder. This class is the bridge between client/context and the
placeOrder method is responsible for deciding what action should to be executed. For example, if the
sell the method should invoke the action in the
StockTrader. The following UML diagram shows the scenario that I have just described.
Agent codes are the following ones:
The most relevant code smell is the
placeOrder method which is coupled to the actions/commands from
StockTrade. There are different techniques to avoid this code smell.In this case, the
Command pattern is a good solution, since we want to log the command’s history.
StockTrade class is the following one:
The result obtained is shown in the following image:
Command pattern — Example 1: A Stock Market — Solution
The idea to decouple the commands from the
Agent class is to create a set of classes for each command. However, the commands share a common interface which allows us to execute the action depending on each concrete command.
That is the reason why we have created the
Order abstract class which will have an abstract method called
execute. This method is the one that will be invoked from the
Agent class (the invoker). Furthermore,
Agent class will have a list of commands to obtain the command’s history.
This way, the agent delegates the responsibility of knowing which operation has to be executed on the object it receives. The main change is that
Agent class will no longer receive a primitive attribute as a parameter (string), since this has no semantic value. Instead, the
Agent class will now receive a command object as a parameter, which provides semantic value.
The new UML diagram using the command pattern is shown below:
The code associate to the
client is the following one:
In this case each
order receives the
StockTrade using DI (Dependency Injection). The
Agent invokes the command using the
placeOrder method, which performs the operation though the
The code associated with the
Agent is the following one:
You may note that the
if-elseif-else control structure is avoided by using the
order.execute method, which delegates the responsibility to each command.
The code associated to the
Order and each order are the following ones:
StockTrade class is not modified in this command. So, the result after these modifications in the execution of the program is shown in the following image:
npm run example1-problem
npm run example1-command-solution1
Another interesting example which is resolved using command pattern is when there are several commands to execute for a robot.
For example, a set of commands as SaveSecret, Clean and Move are asked to a famous robot, R2D2. In the following UML diagram you can see this situation:
The code associated to the clients is the following:
In this example, there are three commands (saveSecretCommand, cleanCommand and moveCommand), two services (StoreService and R2D2Service) and an Agent (R2D2).
The Agent invokes the orders using the
executeCommand method which receives two arguments: 1) The command; 2) The parameters to carry out the previous command.
Therefore, the code associated to the R2D2 is the following one:
R2D2 has a list of commands, which may be listed through the
listCommands method, and stored using the
commands data-structure. Finally, the
executeCommand method is responsible for invoking the execute method of each command.
So, the next step is to create the code associated to the command (abstract class) and each concrete command:
Finally, each command invokes the service responsible for the action, in this case we have used two different services to show that not all the commands delegate responsibility to the same service or class.
The result obtained is shown in the following image:
I have created a npm scripts that run the example shown here after applying the command pattern.
npm run example2-command-solution-1
The command pattern can avoid complexity in your projects because you encapsules the commands in specific class which can be added/removed or changed in any moment (including execution-time).
The most important thing is not implementing the pattern as I have shown you, but to be able to recognise the problem which this specific pattern can resolve, and when you may or may not implement said pattern. This is crucial, since implementation will vary depending on the programming language you use.
The GitHub branch of this post is https://github.com/Caballerog/blog/tree/master/command-pattern