sergeykolbasov / slicker   0.3

Apache License 2.0 GitHub

Putting Slick into repository

Scala versions: 2.11 2.12

Slicker Maven Central Build Status

Using Slick in repository manner.

Repository vs DAO

DAO or Data Access Object is a quick and nice solution to access and represent database inside of your code. Close to the tables, persistence included. Sounds like a silver bullet.
But what if inside of your business domain you want to abstract over data mapping and the way your entity is built? May be you have multiple database sources and don't want to care managing sources on the business level.

Here comes the repository layer which allows to hide details of database access and get rid of duplication of query code.

You could read more about both patterns here:

Point of using Slicker

Slick TableQuery in fact is an implementation of DAO staying close to SQL and exposing SQL logic. Slicker doesn't reject this approach but hides usage of Slick tables under the hood providing sufficient abstraction for simple CRUD operations but still with possibility of doing manual work.


First add following lines to your build.sbt:

libraryDependencies ++= Seq(
    "com.github.imliar" %% "slicker-core" % "0.3",
    "com.github.imliar" %% "slicker-postgres" % "0.3"

Then define table and entity that we're going to use:

import io.slicker.postgres.PostgresDriver.api._
import io.slicker.postgres.{SimpleRecordTable, TableWithId}
import slick.lifted.ProvenShape

case class User(id: Option[Long], name: String)

class UserTable extends SimpleRecordTable[Long, User, Users] {

  override val tableQuery: TableQuery[Users] = TableQuery[Users]


class Users(tag: Tag) extends TableWithId[Long, User](tag, "users") {

  override def id: Rep[Long] = column[Long]("id", O.AutoInc, O.PrimaryKey)

  def name: Rep[String] = column[String]("name")

  override def * : ProvenShape[User] = (id.?, name) <> (User.tupled, User.unapply)


UserTable will be used by Slicker Repository to convert business <> database entities and operate with TableQuery. In this case business model is equal to database one, so it's enough to use SimpleRecordTable.
TableWithId defines id column with provided type so Repository could operate with entity id.

Now define repository interface as a next step. It's not required but just a good practice to split intention and implementation.

import io.slicker.core.Repository

trait UsersRepository extends Repository[Long, User] {
  def findOneByName(name: String): ReadAction[Option[User]]


Except manually provided findAllByName method Repository interface also defines useful everyday methods.
As a result you will get Slick DBIOAction. Running of action is up to user.

Finally define a repository implementation:

import io.slicker.postgres.PostgresDriver.api._
import io.slicker.postgres.PostgreSQLRepository

import scala.concurrent.ExecutionContext

class UsersRepositoryImpl extends PostgreSQLRepository(new UserTable) with UsersRepository {

  //just an example. you could provide your own EC
  override protected implicit val ec: ExecutionContext =

  override def findOneByName(name: String): ReadAction[Option[User]] = findOneBy( === name)

findOneBy is one of methods that come with PostgreSQLRepository and supposed to provide easy-to-use CRUD operations.


Detailed documentation could be found here.


  • Only PostgreSQL implementation available. It'll be easy to support other DBs creating new subproject.