REST API with Go, Chi, MySQL and sqlx

7 minute read

This is a continuation of an earlier post REST API with Go, Chi and InMemory Store. In this tutorial I will extend the service to store data in a MySQL database. I will use Docker to run MySQL and run database migrations.

Project Setup

I will start by copying the content of https://github.com/kashifsoofi/blog-code-samples/tree/main/movies-api-with-go-chi-and-memory-store, placing it in a new folder movies-api-with-go-chi-and-mysql and updating module name in go.mod to match the new folder and updating in source files where its used. This will usually be the root of your git repo and will not be as elaborate as this.

Setup Database Server

I will be using a docker-compose to run MySQL in a docker container. This would allow us the add more services that our rest api is depenedent on e.g. redis server for distributed caching.

Let’s start by adding a new file named as docker-compose.dev-env.yml, feel free to name it as you like. Add following content to add a database instance for movies rest api.

version: '3.7'

services:
  movies.db:
    image: mysql:5.7
    environment:
      - MYSQL_ROOT_PASSWORD=Password123
      - MYSQL_DATABASE=moviesdb
    volumes:
      - moviesdbdata:/var/lib/mysql
    ports:
      - "3306:3306"
    healthcheck:
      test: "mysql -uroot -pPassword123 moviesdb -e 'select 1'"
      timeout: 20s
      interval: 10s
      retries: 10

volumes:
  moviesdbdata:

Open a terminal at the root of the solution where docker-compose file is located and execute following command to start database server.

docker-compose -f docker-compose.dev-env.yml up -d

Database Migrations

Before we can start using MySQL we need to create a table to store our data. I will be using excellent migrate database migrations tool, it can also be imported as a libraray.

For migrations I have created a folder db and another folder named migrations under db. I executed following commands to create migrations.

migrate create -ext sql -dir db/migrations -seq schema_movies_create
migrate create -ext sql -dir db/migrations -seq table_movies_create

This would create 4 files, for each migration there would be an up and a down script, up would be executed when applying migration and down would be executed when rolling back a change.

000001_schema_movies_create.up.sql
```
CREATE SCHEMA IF NOT EXISTS Movies;
```
000001_schema_movies_create.down.sql
```
DROP SCHEMA IF EXISTS Movies;
```

000002_table_movies_create.up.sql

CREATE TABLE IF NOT EXISTS Movies (
  Id          CHAR(36)        NOT NULL UNIQUE,
  Title       VARCHAR(100)    NOT NULL,
  Director    VARCHAR(100)    NOT NULL,
  ReleaseDate DATETIME        NOT NULL,
  TicketPrice DECIMAL(12, 4)  NOT NULL,
  CreatedAt   DATETIME        NOT NULL,
  UpdatedAt   DATETIME        NOT NULL,
  PRIMARY KEY (Id)
) ENGINE=INNODB;

000002_table_movies_create.down.sql
```
DROP TABLE IF EXISTS Movies;
```

I usually create a container that has all database migrations and tool to execute those migrations. Dockerfile to run database migrations is as follows

FROM migrate/migrate

# Copy all db files
COPY ./migrations /migrations

ENTRYPOINT [ "migrate", "-path", "/migrations", "-database"]
CMD ["mysql://root:Password123@tcp(movies.db:3306)/moviesdb up"]

Add following in docker-compose.dev-env.yml file to add migrations container and run migrations on startup. Please remember if you add new migrations, you would need to delete container and movies.db.migrations image to add new migration files in the image.

  movies.db.migrations:
    depends_on:
      movies.db:
        condition: service_healthy
    image: movies.db.migrations
    build:
      context: ./db/
      dockerfile: Dockerfile
    command: "'mysql://root:Password123@tcp(movies.db:3306)/moviesdb' up"

Open a terminal at the root of the project where docker-compose file is location and execute following command to start database server and apply migrations to create Movies schema and Movies table.

docker-compose -f docker-compose.dev-env.yml up -d

MySQL Movies Store

I will be using sqlx to execute queries and map columns to struct fields and vice versa, sqlx is a library which provides a set of extensions on go’s standard database/sql library.

Add a new file named mysql_movies_store.go. Add a new struct MySqlMoviesStore containing databaseUrl and a pointer to sqlx.DB, also add helper methods to connect to database and close connection as well. Also note that I have added a noOpMapper method and set as MapperFunc of sqlx.DB, reason for this is to use the same casing as the struct field name. Default behaviour for sqlx is to map field names to lower case column names.

package store

import (
	"context"
	"database/sql"
	"strings"
	"time"

	_ "github.com/go-sql-driver/mysql"
	"github.com/google/uuid"
	"github.com/jmoiron/sqlx"
)

const driverName = "mysql"

type MySqlMoviesStore struct {
	databaseUrl string
	dbx         *sqlx.DB
}

func NewMySqlMoviesStore(databaseUrl string) *MySqlMoviesStore {
	return &MySqlMoviesStore{
		databaseUrl: databaseUrl,
	}
}

func noOpMapper(s string) string { return s }

func (s *MySqlMoviesStore) connect(ctx context.Context) error {
	dbx, err := sqlx.ConnectContext(ctx, driverName, s.databaseUrl)
	if err != nil {
		return err
	}

	dbx.MapperFunc(noOpMapper)
	s.dbx = dbx
	return nil
}

func (s *MySqlMoviesStore) close() error {
	return s.dbx.Close()
}

Add db tag

Update Movie struct in movies_store.go file to add db tag for ID field, this allows sqlx to map ID field to correct column. Alternative to this is to use the AS in select queries or rename column name in database as ID. All other fields will be mapped correctly by using noOpMapper from the above section.

type Movie struct {
	ID          uuid.UUID `db:"Id"`
	...
}

Context

We did not make use of the Context in the earlier sample movies-api-with-go-chi-and-memory-store, now that we are connecting to an external storage and package we are going to use to run queries support methods accepting Context we will update our store.Interface to accept Context and use that when running queries. store.Interface will be updated as follows

type Interface interface {
	GetAll(ctx context.Context) ([]Movie, error)
	GetByID(ctx context.Context, id uuid.UUID) (Movie, error)
	Create(ctx context.Context, createMovieParams CreateMovieParams) error
	Update(ctx context.Context, id uuid.UUID, updateMovieParams UpdateMovieParams) error
	Delete(ctx context.Context, id uuid.UUID) error
}

We will also need to update MemoryMoviesStore methods to accept Context to satisfy store.Interface and update methods in movies_handler to pass request context using r.Context() when calling store methods.

Create

We connect to database using connect helper method, create a new instance of Movie and execute insert query with NamedExecContext. We are handling an error and return DuplicateKeyError if returned error contains text Error 1062. If insert is successful then we return nil. Create function looks like

func (s *MySqlMoviesStore) Create(ctx context.Context, createMovieParams CreateMovieParams) error {
	err := s.connect(ctx)
	if err != nil {
		return err
	}
	defer s.close()

	movie := Movie{
		ID:          createMovieParams.ID,
		Title:       createMovieParams.Title,
		Director:    createMovieParams.Director,
		ReleaseDate: createMovieParams.ReleaseDate,
		TicketPrice: createMovieParams.TicketPrice,
		CreatedAt:   time.Now().UTC(),
		UpdatedAt:   time.Now().UTC(),
	}

	if _, err := s.dbx.NamedExecContext(
		ctx,
		`INSERT INTO Movies
			(Id, Title, Director, ReleaseDate, TicketPrice, CreatedAt, UpdatedAt)
		VALUES
			(:Id, :Title, :Director, :ReleaseDate, :TicketPrice, :CreatedAt, :UpdatedAt)`,
		movie); err != nil {
		if strings.Contains(err.Error(), "Error 1062") {
			return &DuplicateKeyError{ID: createMovieParams.ID}
		}
		return err
	}

	return nil
}

GetAll

We connect to database using connect helper method, then use SelectContext method of sqlx to execute query, sqlx would map the columns to fields. If query is successful then we return the slice of loaded movies.

func (s *MySqlMoviesStore) GetAll(ctx context.Context) ([]Movie, error) {
	err := s.connect(ctx)
	if err != nil {
		return nil, err
	}
	defer s.close()

	var movies []Movie
	if err := s.dbx.SelectContext(
		ctx,
		&movies,
		`SELECT
			Id, Title, Director, ReleaseDate, TicketPrice, CreatedAt, UpdatedAt
		FROM Movies`); err != nil {
		return nil, err
	}

	return movies, nil
}

If there is an error in parsing DATETIME column, remember to add parseTime=true parameter to your connection string.

GetByID

We connect to database using connect helper method, then use GetContext method to execute select query, sqlx would map the columns to fields. If the driver returns sql.ErrNoRows then we return store.RecordNotFoundError. If successful loaded movie record is returned.

func (s *MySqlMoviesStore) GetByID(ctx context.Context, id uuid.UUID) (Movie, error) {
	err := s.connect(ctx)
	if err != nil {
		return Movie{}, err
	}
	defer s.close()

	var movie Movie
	if err := s.dbx.GetContext(
		ctx,
		&movie,
		`SELECT
			Id, Title, Director, ReleaseDate, TicketPrice, CreatedAt, UpdatedAt
		FROM Movies
		WHERE Id = ?`,
		id); err != nil {
		if err != sql.ErrNoRows {
			return Movie{}, err
		}

		return Movie{}, &RecordNotFoundError{}
	}

	return movie, nil
}

Update

We connect to database using connect helper method, then use NamedExecContext method to execute query to update an existing record.

func (s *MySqlMoviesStore) Update(ctx context.Context, id uuid.UUID, updateMovieParams UpdateMovieParams) error {
	err := s.connect(ctx)
	if err != nil {
		return err
	}
	defer s.close()

	movie := Movie{
		ID:          id,
		Title:       updateMovieParams.Title,
		Director:    updateMovieParams.Director,
		ReleaseDate: updateMovieParams.ReleaseDate,
		TicketPrice: updateMovieParams.TicketPrice,
		UpdatedAt:   time.Now().UTC(),
	}

	if _, err := s.dbx.NamedExecContext(
		ctx,
		`UPDATE Movies
		SET Title = :Title, Director = :Director, ReleaseDate = :ReleaseDate, TicketPrice = :TicketPrice, UpdatedAt = :UpdatedAt
		WHERE Id = :Id`,
		movie); err != nil {
		return err
	}

	return nil
}

Delete

We connect to database using connect helper method, then execute query to delete an existing record using ExecContext.

func (s *MySqlMoviesStore) Delete(ctx context.Context, id uuid.UUID) error {
	err := s.connect(ctx)
	if err != nil {
		return err
	}
	defer s.close()

	if _, err := s.dbx.ExecContext(
		ctx,
		`DELETE FROM Movies
		WHERE id = ?`, id); err != nil {
		return err
	}

	return nil
}

Database Configuration

Add a new struct named Database in config.go and add that to Configuration struct as well.

type Configuration struct {
	HTTPServer
	Database
}
...
type Database struct {
	DatabaseURL        string `envconfig:"DATABASE_URL" required:"true"`
	LogLevel           string `envconfig:"DATABASE_LOG_LEVEL" default:"warn"`
	MaxOpenConnections int    `envconfig:"DATABASE_MAX_OPEN_CONNECTIONS" default:"10"`
}

Dependency Injection

Update main.go as follows to create a new instance of MySqlMoviesStore, I have opted to create instance of MySqlMoviesStore instead of MemoryMoviesStore, solution can be enhanced to create either one of the dependency based on a configuration.

// store := store.NewMemoryMoviesStore()
store := store.NewMySqlMoviesStore(cfg.DatabaseURL)

Test

I am not adding any unit or integration tests for this tutorial, perhaps a following tutorial. But all the endpoints can be tested either using Postman for by following test plan from previous article.

You can start rest api with mysql running in docker by executing following

DATABASE_URL=root:Password123@tcp(localhost:3306)/moviesdb?parseTime=true go run main.go

Source

Source code for the demo application is hosted on GitHub in blog-code-samples repository.

References

In no particular order

Twitter Facebook LinkedIn

Your email address will not be published. Required fields are marked *

Kashif Soofi

REST API with Go, Chi, MySQL and sqlx

Project Setup

Setup Database Server

Database Migrations

MySQL Movies Store

Add db tag

Context

Create

GetAll

GetByID

Update

Delete

Database Configuration

Dependency Injection

Test

Source

References

Leave a Comment

You May Also Enjoy

GTK4 `dotnet new` Project Template

GTK4 ColumnView with .NET 8

GTK4 GridView with .NET 8

GTK4 ListView with .NET 8

Kashif Soofi

Project Setup

Setup Database Server

Database Migrations

MySQL Movies Store

Add db tag

Context

Create

GetAll

GetByID

Update

Delete

Database Configuration

Dependency Injection

Test

Source

References

Leave a Comment

You May Also Enjoy

GTK4 dotnet new Project Template

GTK4 ColumnView with .NET 8

GTK4 GridView with .NET 8

GTK4 ListView with .NET 8

GTK4 `dotnet new` Project Template