After recently joining Central Tech, I was given my first project: implement a new product creation and querying API in Go. I took this as a chance for me to finally dive deep into Go, something I've been wanted to do after having heard praises about it from many others. I also decided to start detailing the things I learned along the way, both for personal reference and to ensure that I understand concepts enough to explain it in writing. As such, this article and will hopefully come to act as a resource that I can refer back to, and hopefully be helpful to someone else as well. my site Basics This piece assumes that you have go installed and working on your computer. If not, please see the Go language's official getting started . guide Before writing the actual functions to execute, we first need to define our program and import the necessary packages. Create a new file called and add the following lines. main.go main ( ) package import "log" "net/http" Let's go over each of the components of the code above. Packages are Go's source code, organized into system directories. They allow for code reusability across different applications. In order to differentiate an executable program from a library to be used by other programs, Go dictates that we includes in the header of the main file of any executables. When compiling, that tells the Go compiler that the package should compile as an executable program rather than being built as a library. Packages package main package main Imports The keyword is used for importing packages into our program or another package. When importing packages, the Go compiler will look in the locations specified by the two environment variables: import where packages from the shipped with Go are stored $GOROOT standard library for fetched third-party or self-made packages. $GOPATH Note: Go by design does not allow declaration of unused import and will adamantly complain about it during compile time if any is found In our case, both packages comes from the standard library: allows us to log errors and other issues. log gives us HTTP client and server implementation for building the actual API net/http After declaring the packages, we can begin defining and serving our API. Without using third party libraries and routers, there are two main 'approaches' to this. If we implement our handler as a function, we'd use . Otherwise, if we'd implemented our handler as a type with a method, we'd use . http.HandleFunc ServeHTTP http.Handle For simple implementations, the first option is much easier and clearer to read and understand. But using a struct type allows storing of useful information in it, an example of which is the from the standard library. The struct contains the root directory for file service. See this for a more detailed explanation. file server Stack Overflow post As our program is rather simple and straightforward, we'll stick to the first approach. Note: HTTP handlers are processes that runs in response to a request made to a web application. For more information, see its Wikipedia page Building a Web Server Now we get to defining how our API server will actually behave, and the content we're going to send as our response. Add the following lines below our imports and package declarations: {
  w.WriteHeader(http.StatusOK)
  w.Header().Set( , )
  w.Write([] ( ))
} {
  http.HandleFunc( ,apiResponse)
  log.Fatal(http.ListenAndServe( , ))
} func apiResponse (w http.ResponseWriter, r *http.Request) "Content-Type" "application/json" byte `{"message":"hello world!"}` func main () "/" ":8080" nil Again, let's run through each function. apiResponse The function apiResponse is responsible for writing the response code and JSON. The function itself takes two arguments: of type assembles our HTTP server's responses, while reads and parse the requests. Now to go through each line of the function: w http.ResponseWriter r - writes the HTTP 200, indicating that the requested resource (the JSON response in our case) has been successfully fetched and is transmitted in the message body. w.WriteHeader(http.StatusOK) response code - indicates the return type of the requested resource. w.Header().Set("Content-Type", "application/json") - finally writes our response in the message body. w.Write([]byte('{"message":"hello world!"}')) main The function serves a special purpose in Go: it acts as the entry point of the executable programs. It does not take any argument nor return anything. And unlike some other languages, we do not need to explicitly call , as Go automatically calls the function when run. main() main In our main function, the call tells the HTTP package to handle all requests to the web root ("/") with our apiResponse function. It then calls , specifying that it should listen on port on any interface. HandleFunc http.ListenAndServe 8080 If we now run the command and navigate to on our web browser, we should see something like this. go run . localhost:8080 We can also enter the same URL into our REST Client (i.e. , , ) using the "GET" method to get the same result. Postman Insomnia Paw Making it RESTful Now that our server can return a response, we can configure it to do different things depending on the method being requested. To do so, we can modify our function as such: apiResponse { w.Header().Set( , ) r.Method { :
      w.WriteHeader(http.StatusOK)
      w.Write([] ( )) :
        w.WriteHeader(http.StatusCreated)
        w.Write([] ( )) :
        w.WriteHeader(http.StatusNotFound)
        w.Write([] ( ))
    }
} func apiResponse (w http.ResponseWriter, r *http.Request) // Set the return Content-Type as JSON like before "Content-Type" "application/json" // Change the response depending on the method being requested switch case "GET" byte `{"message": "GET method requested"}` case "POST" byte `{"message": "POST method requested"}` default byte `{"message": "Can't find method requested"}` Now if we close the previous version of our server with and start it once again, we should get different responses depending on the method we requested from our REST Client. ctrl-c Our final should look something like: main.go main ( ) { w.Header().Set( , ) r.Method { :
      w.WriteHeader(http.StatusOK)
      w.Write([] ( )) :
        w.WriteHeader(http.StatusCreated)
        w.Write([] ( )) :
        w.WriteHeader(http.StatusNotFound)a
        w.Write([] ( ))
    }
} {
  http.HandleFunc( ,apiResponse)
  log.Fatal(http.ListenAndServe( , ))
} package import "log" "net/http" func apiResponse (w http.ResponseWriter, r *http.Request) // Set the return Content-Type as JSON like before "Content-Type" "application/json" // Change the response depending on the method being requested switch case "GET" byte `{"message": "GET method requested"}` case "POST" byte `{"message": "POST method requested"}` default byte `{"message": "Can't find method requested"}` func main () "/" ":8080" nil Closing This covers the basic of building a REST API server using Go. In later posts I plan on going over the library in more detail, as well as using third-party router libraries such as and for more complex routings. net/http gorillia/mux go-chi/chi If you have any questions, feel free to reach out to me on or via . Twitter email

Twitter

Read My Stories

Too Long; Didn't Read

In your car, at home, or at work — Bosch technology shapes many areas of life.

Building a Simple Rest API with Go

Too Long; Didn't Read

Company Mentioned

Sawit Trisirisatayawong

About Author

TOPICS

THIS ARTICLE WAS FEATURED IN...

Building a Simple Rest API with Go

Too Long; Didn't Read

Company Mentioned

Sawit Trisirisatayawong

About Author

TOPICS

THIS ARTICLE WAS FEATURED IN...

RELATED STORIES