ksql/readme.template.md

KSQL the Keep it Simple SQL library

KSQL was created to offer an actually simple and satisfactory tool for interacting with SQL Databases in Golang.

The core goal of KSQL is not to offer new features that are unavailable on other libraries (although we do have some), but to offer a well-thought and well-planned API so that users have an easier time, learning, debugging, and avoiding common pitfalls.

KSQL is also decoupled from its backend so that the actual communication with the database is performed by well-known and trusted technologies, namely: pgx and database/sql. You can even create your own backend adapter for KSQL which is useful in some situations.

In this README you will find examples for "Getting Started" with the library, for more advanced use-cases please read our Wiki.

Let's start with some Code:

This short example below is a TLDR version for illustrating how easy it is to use KSQL.

You will find more complete examples on the sections below.

package main

import (
	"context"
	"fmt"
	"log"
	"os"

	"github.com/vingarcia/ksql"
	"github.com/vingarcia/ksql/adapters/kpgx"
)

var UsersTable = ksql.NewTable("users", "user_id")

type User struct {
	ID   int    `ksql:"user_id"`
	Name string `ksql:"name"`
	Type string `ksql:"type"`
}

func main() {
	ctx := context.Background()
	db, err := kpgx.New(ctx, os.Getenv("POSTGRES_URL"), ksql.Config{})
	if err != nil {
		log.Fatalf("unable connect to database: %s", err)
	}
	defer db.Close()

	// For querying only some attributes you can
	// create a custom struct like this:
	var count []struct {
		Count string `ksql:"count"`
		Type string `ksql:"type"`
	}
	err = db.Query(ctx, &count, "SELECT type, count(*) as count FROM users WHERE type = $1 GROUP BY type", "admin")
	if err != nil {
		log.Fatalf("unable to query users: %s", err)
	}

	fmt.Println("number of users by type:", count)

	// For loading entities from the database KSQL can build
	// the SELECT part of the query for you if you omit it like this:
	var users []User
	err = db.Query(ctx, &users, "FROM users WHERE type = $1", "admin")
	if err != nil {
		log.Fatalf("unable to query users: %s", err)
	}

	fmt.Println("users:", users)
}

We currently have 4 constructors available, one of them is illustrated above (kpgx.New()), the other ones have the exact same signature but work on different databases, they are:

• kpgx.New(ctx, os.Getenv("DATABASE_URL"), ksql.Config{}) for Postgres, it works on top of pgxpool
• kmysql.New(ctx, os.Getenv("DATABASE_URL"), ksql.Config{}) for MySQL, it works on top of database/sql
• ksqlserver.New(ctx, os.Getenv("DATABASE_URL"), ksql.Config{}) for SQLServer, it works on top of database/sql
• ksqlite3.New(ctx, os.Getenv("DATABASE_URL"), ksql.Config{}) for SQLite3, it works on top of database/sql

The KSQL Interface

The current interface contains the methods the users are expected to use, and it is also used for making it easy to mock the whole library if needed.

This interface is declared in the project as ksql.Provider and is displayed below.

We plan on keeping it very simple with a small number of well thought functions that cover all use-cases, so don't expect many additions:

// Provider describes the KSQL public behavior
//
// The Insert, Patch, Delete and QueryOne functions return `ksql.ErrRecordNotFound`
// if no record was found or no rows were changed during the operation.
type Provider interface {
	Insert(ctx context.Context, table Table, record interface{}) error
	Patch(ctx context.Context, table Table, record interface{}) error
	Delete(ctx context.Context, table Table, idOrRecord interface{}) error

	Query(ctx context.Context, records interface{}, query string, params ...interface{}) error
	QueryOne(ctx context.Context, record interface{}, query string, params ...interface{}) error
	QueryChunks(ctx context.Context, parser ChunkParser) error

	Exec(ctx context.Context, query string, params ...interface{}) (Result, error)
	Transaction(ctx context.Context, fn func(Provider) error) error
}

Using KSQL

In the example below we'll cover all the most common use-cases such as:

1. Inserting records
2. Updating records
3. Deleting records
4. Querying one or many records
5. Making transactions

More advanced use cases are illustrated on their own pages on our Wiki:

• Querying in Chunks for Big Queries
• Avoiding Code Duplication with the Select Builder
• Reusing Existing Structs on Queries with JOINs
• Testing Tools and ksql.Mock

For the more common use-cases please read the example below, which is also available here if you want to compile it yourself.

{{ .crudExample -}}

Benchmark Comparison

The results of the benchmark are good: they show that KSQL is in practical terms, as fast as sqlx which was our goal from the start.

To understand the benchmark below you must know that all tests are performed using Postgres 12.1 and that we are comparing the following tools:

• KSQL using the adapter that wraps database/sql
• KSQL using the adapter that wraps pgx
• database/sql
• sqlx
• pgx (with pgxpool)
• gorm
• sqlc
• sqlboiler

For each of these tools we are running 3 different queries:

The insert-one query looks like:

INSERT INTO users (name, age) VALUES ($1, $2) RETURNING id

The single-row query looks like:

SELECT id, name, age FROM users OFFSET $1 LIMIT 1

The multiple-rows query looks like:

SELECT id, name, age FROM users OFFSET $1 LIMIT 10

Keep in mind that some of the tools tested (like GORM) actually build the queries internally so the actual code used for the benchmark might differ a little bit from the example ones above.

Without further ado, here are the results:

$ make bench TIME=5s
{{ .benchmark -}}

Running the KSQL tests (for contributors)

The tests use docker-test for setting up all the supported databases, which means that:

• You need to have docker installed

• You must be able to run docker without sudo, i.e. if you are not root you should add yourself to the docker group, e.g.:

  $ sudo usermod <your_username> -aG docker
  

  And then restart your login session (or just reboot)

After that you can just run the tests by using:

make test

But it is recommended to first download the required images using:

docker pull postgres:14.0
docker pull mysql:8.0.27
docker pull mcr.microsoft.com/mssql/server:2017-latest

Otherwise the first attempt to run the tests will spend a long time downloading these images and then fail because the TestMain() function is configured to kill the containers after 20 seconds.

TODO List

• Update ksqltest.FillStructWith to work with ksql:"..,json" tagged attributes
• Improve error messages (ongoing)

Optimization Oportunities

• Test if using a pointer on the field info is faster or not
• Consider passing the cached structInfo as argument for all the functions that use it, so that we don't need to get it more than once in the same call.
• Use a cache to store often used queries (like pgx)
• Preload the insert method for all dialects inside ksql.NewTable()
• Use prepared statements for the helper functions, Update, Insert and Delete.

Features for a possible V2

• Change the .Transaction(db ksql.Provider) to a .Transaction(ctx context.Context)
• Make the .Query() method to return a type Query interface { One(); All(); Chunks(); }
• Have an Update() method that updates without ignoring NULLs as Patch() does
  • Have a new Modifier skipNullUpdates so that the Update function will do the job of the Patch
  • Remove the Patch function.
• Rename NewTable() to just Table() so it feels right to declare it inline when convenient