浅谈Slick（4）－ Slick301：我的Slick开发项目设置

  前面几篇介绍里尝试了一些Slick的功能和使用方式，看来基本可以满足用scala语言进行数据库操作编程的要求，而且有些代码可以通过函数式编程模式来实现。我想，如果把Slick当作数据库操作编程主要方式的话，可能需要先制定一套比较规范的模式来应付日常开发（也要考虑团队开发）、测试和维护。首先从项目结构来说，我发现由Intellij-Idea IDE界面直接产生的SBT项目结构已经比较理想了。在src/main/resources是scala项目获取配置文件的默认目录、我们可以按照需要在src/main/scala下增加代码子目录（package）及在src/main/test下摆放测试代码。配置文件application.conf、logback.xml是放在src/main/resources下的。application.conf是Slick的配置文件，logback.xml是跟踪器logback（log4j）的配置文件。Slick把jdbc api集成到scala编程语言里，能够支持多种数据库。也就是说Slick提供了多种数据库的驱动api。Slick支持在配置文件application.conf里配置数据库功能模式，这样我们就可以在正式部署软件时才通过修订application.conf里的配置来决定具体的数据库种类和参数。当然前提是我们的程序代码不能依赖任何特别的数据库api。我们从表结构设定开始，先看看上篇Slick101里的例子：

package com.datatech.learn.slick101
import slick.driver.H2Driver.api._
object slick101 {

/* ----- schema  */
  //表字段对应模版
  case class AlbumModel (id: Long
                   ,title: String
                   ,year: Option[Int]
                   ,artist: String
                   )
  //表结构: 定义字段类型, * 代表结果集字段
  class AlbumTable(tag: Tag) extends Table[AlbumModel](tag, "ALBUMS") {
    def id = column[Long]("ID",O.AutoInc,O.PrimaryKey)
    def title = column[String]("TITLE")
    def year = column[Option[Int]]("YEAR")
    def artist = column[String]("ARTIST",O.Default("Unknown"))
    def * = (id,title,year,artist) <> (AlbumModel.tupled, AlbumModel.unapply)
  }
  //库表实例
  val albums = TableQuery[AlbumTable]

我们可以看到这段代码依赖了slick.driver.H2Driver.api，是专门针对H2 Database的了。我们可以用依赖注入（dependency injection, IOC）来解决这个依赖问题。先试试用最传统的依赖注入方式：传入参数来注入这个数据库驱动依赖，把代码放在src/main/scala/model/TableDefs.scala里：

package com.bayakala.learn.slick301.model
import slick.driver.JdbcProfile
class TableDefs(val dbDriver: JdbcProfile) {
  import dbDriver.api._
  case class Supplier(id: Long
                      , name: String
                      , contact: Option[String]
                      , website: Option[String])
  final class Suppliers(tag: Tag) extends Table[Supplier](tag,"SUPPLERS") {
    def id = column[Long]("ID",O.AutoInc,O.PrimaryKey)
    def name = column[String]("NAME")
    def contact = column[Option[String]]("CONTACT")
    def website = column[Option[String]]("WEBSITE")
    def * = (id, name, contact, website) <> (Supplier.tupled,Supplier.unapply)
    def nidx = index("NM_IDX",name,unique = true)
  }
  val suppliers = TableQuery[Suppliers]

  case class Coffee(id: Long
                     ,name: String
                     ,supid: Long
                     ,price: Double
                     ,sales: Int)
  final class Coffees(tag: Tag) extends Table[Coffee](tag, "COFFEES") {
    def id = column[Long]("ID",O.AutoInc,O.PrimaryKey)
    def name = column[String]("NAME")
    def supid = column[Long]("SUPID")
    def price = column[Double]("PRICE",O.Default(0.0))
    def sales = column[Int]("SALES",O.Default(0))
    def * = (id,name,supid,price,sales) <> (Coffee.tupled, Coffee.unapply)
    def fk_sup = foreignKey("FK_SUP",supid,suppliers)(_.id,onDelete = ForeignKeyAction.Restrict,onUpdate = ForeignKeyAction.Cascade)
    def supidx = index("SUP_IDX",supid,unique = false)
    def nidx = index("NM_IDX",name,unique = true)
  }
  val coffees = TableQuery[Coffees]

}

注意我们是把JdbcProfile作为参数注入了class TableDefs里。如果TableDefs经常需要作为其它类的父类继承的话，设计成trait能更加灵活的进行类型混合（type mixing）。这样的需求可以用cake pattern方式进行依赖注入。我们在需要src/main/scala/config/AppConfig.scala里定义依赖界面trait DBConfig：

package com.bayakala.learn.slick301.config
import slick.driver.JdbcProfile
trait DBConfig {
  val jdbcDriver: JdbcProfile
  import jdbcDriver.api._
  val db: Database
}

后面我们可以通过实现多种DBConfig实例方式来构建开发、测试、部署等数据库环境。为了方便示范，我们设计几个基本的Query Action，放在src/main/scala/access/DAOs.scala里，用cake pattern注入依赖DBConfig：

package com.bayakala.learn.slick301.access
import com.bayakala.learn.slick301.config
import com.bayakala.learn.slick301.config.DBConfig
import com.bayakala.learn.slick301.model.TableDefs
trait DAOs { dbconf: DBConfig =>
  import jdbcDriver.api._
  //注入依赖
  val tables = new TableDefs(dbconf.jdbcDriver)
  import tables._
  //suppliers queries
  val createSupplierTable = suppliers.schema.create
  val allSuppliers = suppliers.result
  def insertSupplier(id:Long,name:String,address:Option[String],website:Option[String])
  = suppliers += Supplier(id,name,address,website)
  def insertSupbyName(n: String) = suppliers.map(_.name) += n
  //coffees queries
  val createCoffeeTable = coffees.schema.create
  val allCoffees = coffees.result
  def insertCoffee(c: (Long,String,Long,Double,Int)) =
    coffees += Coffee(id=c._1, name=c._2,supid=c._3,price=c._4,sales=c._5)

}

dbconf: DBConfig => 的意思是在进行DAOs的实例化时必须混入（mixing）DBConfig类。

以上两个代码文件TableDefs.scala和DAOs.scala在注入依赖后都能够顺利通过编译了。

我们在src/main/scala/main/Main.scala里测试运算DAOs里的query action：

package com.bayakala.learn.slick301.main
import com.bayakala.learn.slick301.config.DBConfig
import com.bayakala.learn.slick301.access.DAOs

import scala.concurrent.{Await, Future}
import scala.util.{Failure, Success}
import scala.concurrent.duration._
import scala.concurrent.ExecutionContext.Implicits.global
import slick.backend.DatabaseConfig
import slick.driver.{H2Driver, JdbcProfile}
object Main {
 
  object Actions extends DAOs with DBConfig {
    override lazy val jdbcDriver: JdbcProfile = H2Driver
    val dbConf: DatabaseConfig[H2Driver] = DatabaseConfig.forConfig("h2")
    override val db = dbConf.db
  }
  import Actions._
  
  def main(args: Array[String]) = {
    val res = db.run(createSupplierTable).andThen {
      case Success(_) => println("supplier table created")
      case Failure(_) => println("unable to create supplier table")
    }
    Await.ready(res, 3 seconds)

    val res2 = db.run(insertSupbyName("Acme Coffee Co."))
    Await.ready(res2, 3 seconds)

    Await.ready(db.run(allSuppliers), 10 seconds).foreach(println)

    val res10 = db.run(createCoffeeTable).andThen {
      case Success(_) => println("coffee table created")
      case Failure(_) => println("unable to create coffee table")
    }
    Await.ready(res10, 3 seconds)

    val res11 = db.run(insertCoffee((101,"Columbia",1,158.0,0)))
    Await.ready(res11, 3 seconds)

    Await.ready(db.run(allCoffees), 10 seconds).foreach(println)

  }

}

Actions是DAOs的实例。我们看到必须把DBConfig混入（mixin）。但是我们构建的数据库又变成了专门针对H2的api了，这样的话每次变动数据库对象我们就必须重新编译Main.scala，不符合上面我们提到的要求。我们可以把目标数据库放到application.conf里，然后在Main.scala里用typesafe-config实时根据application.conf里的设置确定数据库参数。src/main/resources/application.conf内容如下：

app = {
  dbconfig = h2
}

h2 {
  driver = "slick.driver.H2Driver$"
  db {
    url = "jdbc:h2:~/slickdemo;mv_store=false"
    driver = "org.h2.Driver"
    connectionPool = HikariCP
    numThreads = 10
    maxConnections = 12
    minConnections = 4
    keepAliveConnection = true
  }
}

h2mem = {
  url = "jdbc:h2:mem:slickdemo"
  driver = org.h2.Driver
  connectionPool = disabled
  keepAliveConnection = true
}

mysql {
  driver = "slick.driver.MySQLDriver$"
  db {
    url = "jdbc:mysql://localhost/slickdemo"
    driver = com.mysql.jdbc.Driver
    keepAliveConnection = true
    user="root"
    password="123"
    numThreads=10
    maxConnections = 12
    minConnections = 4
  }
}

mysqldb = {
  dataSourceClass = "com.mysql.jdbc.jdbc2.optional.MysqlDataSource"
  properties {
    user = "root"
    password = "123"
    databaseName = "slickdemo"
    serverName = "localhost"
  }
  numThreads = 10
  maxConnections = 12
  minConnections = 4
}

postgres {
  driver = "slick.driver.PostgresDriver$"
  db {
    url = "jdbc:postgresql://127.0.0.1/slickdemo"
    driver = "org.postgresql.Driver"
    connectionPool = HikariCP
    user = "slick"
    password = "123"
    numThreads = 10
    maxConnections = 12
    minConnections = 4
  }
}

postgressdb = {
  dataSourceClass = "org.postgresql.ds.PGSimpleDataSource"
  properties = {
    databaseName = "slickdemo"
    user = "slick"
    password = "123"
  }
  connectionPool = HikariCP
  numThreads = 10
  maxConnections = 12
  minConnections = 4
}

mssql {
  driver = "com.typesafe.slick.driver.ms.SQLServerDriver$"
  db {
    url = "jdbc:sqlserver://host:port"
    driver = com.microsoft.sqlserver.jdbc.SQLServerDriver
    connectionTimeout = 30 second
    connectionPool = HikariCP
    user = "slick"
    password = "123"
    numThreads = 10
    maxConnections = 12
    minConnections = 4
    keepAliveConnection = true
  }
}

tsql {
  driver = "slick.driver.H2Driver$"
  db = ${h2mem}
}

现在application.conf里除了数据库配置外又加了个app配置。我们在Main.scala里实例化DAOs时可以用typesafe-config读取app.dbconfig值后设定jdbcDriver和db：

  object Actions extends DAOs with DBConfig {
    import slick.util.ClassLoaderUtil
    import scala.util.control.NonFatal
    import com.typesafe.config.ConfigFactory

    def getDbConfig: String =
        ConfigFactory.load().getString("app.dbconfig")

    def getDbDriver(path: String): JdbcProfile = {
      val config = ConfigFactory.load()
      val n = config.getString((if (path.isEmpty) "" else path + ".") + "driver")
      val untypedP = try {
        if (n.endsWith("$")) ClassLoaderUtil.defaultClassLoader.loadClass(n).getField("MODULE$").get(null)
        else ClassLoaderUtil.defaultClassLoader.loadClass(n).newInstance()
      } catch {
        case NonFatal(ex) =>
          throw new SlickException(s"""Error getting instance of Slick driver "$n"""", ex)
      }
      untypedP.asInstanceOf[JdbcProfile]
    }
    
    override lazy val jdbcDriver: JdbcProfile = getDbDriver(getDbConfig)
    val dbConf: DatabaseConfig[JdbcProfile] = DatabaseConfig.forConfig(getDbConfig)
    override val db = dbConf.db
  }

现在我们只需要改变application.conf里的app.dbconfig就可以转换目标数据库参数了。实际上，除了数据库配置，我们还可以在application.conf里进行其它类型的配置。然后用typesafe-config实时读取。如果不想在application.conf进行数据库之外的配置，可以把其它配置放在任何文件里，然后用ConfigFactory.load(path)来读取。

另外，在软件开发过程中跟踪除错也是很重要的。我们可以用logback来跟踪Slick、HikariCP等库的运行状态。logback配置在src/main/resources/logback.xml：

<?xml version="1.0" encoding="UTF-8"?>

<configuration>
    <appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">
        <encoder>
            <pattern>%d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</pattern>
        </encoder>
    </appender>

    <logger name="application" level="DEBUG"/>
    <logger name="com.zaxxer.hikari" level="DEBUG"/>
    <logger name="slick" level="DEBUG"/>

    <root level="DEBUG">
        <appender-ref ref="STDOUT"/>
    </root>
</configuration>

DEBUG值可以显示最详细的状态信息。

好了，我把这次示范代码提供在下面：

build.sbt：

name := "learn-slick301"

version := "1.0"

scalaVersion := "2.11.8"

libraryDependencies ++= Seq(
  "com.typesafe.slick" %% "slick" % "3.1.1",
  "com.h2database" % "h2" % "1.4.191",
  "com.typesafe.slick" %% "slick-hikaricp" % "3.1.1",
  "ch.qos.logback" % "logback-classic" % "1.1.7",
  "org.typelevel" %% "cats" % "0.7.2"

)

src/main/resources/

application.conf:

app = {
  dbconfig = h2
}

h2 {
  driver = "slick.driver.H2Driver$"
  db {
    url = "jdbc:h2:~/slickdemo;mv_store=false"
    driver = "org.h2.Driver"
    connectionPool = HikariCP
    numThreads = 10
    maxConnections = 12
    minConnections = 4
    keepAliveConnection = true
  }
}

h2mem = {
  url = "jdbc:h2:mem:slickdemo"
  driver = org.h2.Driver
  connectionPool = disabled
  keepAliveConnection = true
}

mysql {
  driver = "slick.driver.MySQLDriver$"
  db {
    url = "jdbc:mysql://localhost/slickdemo"
    driver = com.mysql.jdbc.Driver
    keepAliveConnection = true
    user="root"
    password="123"
    numThreads=10
    maxConnections = 12
    minConnections = 4
  }
}

mysqldb = {
  dataSourceClass = "com.mysql.jdbc.jdbc2.optional.MysqlDataSource"
  properties {
    user = "root"
    password = "123"
    databaseName = "slickdemo"
    serverName = "localhost"
  }
  numThreads = 10
  maxConnections = 12
  minConnections = 4
}

postgres {
  driver = "slick.driver.PostgresDriver$"
  db {
    url = "jdbc:postgresql://127.0.0.1/slickdemo"
    driver = "org.postgresql.Driver"
    connectionPool = HikariCP
    user = "slick"
    password = "123"
    numThreads = 10
    maxConnections = 12
    minConnections = 4
  }
}

postgressdb = {
  dataSourceClass = "org.postgresql.ds.PGSimpleDataSource"
  properties = {
    databaseName = "slickdemo"
    user = "slick"
    password = "123"
  }
  connectionPool = HikariCP
  numThreads = 10
  maxConnections = 12
  minConnections = 4
}

mssql {
  driver = "com.typesafe.slick.driver.ms.SQLServerDriver$"
  db {
    url = "jdbc:sqlserver://host:port"
    driver = com.microsoft.sqlserver.jdbc.SQLServerDriver
    connectionTimeout = 30 second
    connectionPool = HikariCP
    user = "slick"
    password = "123"
    numThreads = 10
    maxConnections = 12
    minConnections = 4
    keepAliveConnection = true
  }
}

tsql {
  driver = "slick.driver.H2Driver$"
  db = ${h2mem}
}

logback.xml:

<?xml version="1.0" encoding="UTF-8"?>

<configuration>
    <appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">
        <encoder>
            <pattern>%d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</pattern>
        </encoder>
    </appender>

    <logger name="application" level="DEBUG"/>
    <logger name="com.zaxxer.hikari" level="DEBUG"/>
    <logger name="slick" level="DEBUG"/>

    <root level="DEBUG">
        <appender-ref ref="STDOUT"/>
    </root>
</configuration>

src/main/scala/config/AppConfig.scala:

package com.bayakala.learn.slick301.config
import slick.driver.JdbcProfile
trait DBConfig {
  val jdbcDriver: JdbcProfile
  import jdbcDriver.api._
  val db: Database
}

src/main/scala/model/TableDefs.scala:

package com.bayakala.learn.slick301.model
import slick.driver.JdbcProfile
class TableDefs(val dbDriver: JdbcProfile) {
  import dbDriver.api._
  case class Supplier(id: Long
                      , name: String
                      , contact: Option[String]
                      , website: Option[String])
  final class Suppliers(tag: Tag) extends Table[Supplier](tag,"SUPPLERS") {
    def id = column[Long]("ID",O.AutoInc,O.PrimaryKey)
    def name = column[String]("NAME")
    def contact = column[Option[String]]("CONTACT")
    def website = column[Option[String]]("WEBSITE")
    def * = (id, name, contact, website) <> (Supplier.tupled,Supplier.unapply)
    def nidx = index("NM_IDX",name,unique = true)
  }
  val suppliers = TableQuery[Suppliers]

  case class Coffee(id: Long
                     ,name: String
                     ,supid: Long
                     ,price: Double
                     ,sales: Int)
  final class Coffees(tag: Tag) extends Table[Coffee](tag, "COFFEES") {
    def id = column[Long]("ID",O.AutoInc,O.PrimaryKey)
    def name = column[String]("NAME")
    def supid = column[Long]("SUPID")
    def price = column[Double]("PRICE",O.Default(0.0))
    def sales = column[Int]("SALES",O.Default(0))
    def * = (id,name,supid,price,sales) <> (Coffee.tupled, Coffee.unapply)
    def fk_sup = foreignKey("FK_SUP",supid,suppliers)(_.id,onDelete = ForeignKeyAction.Restrict,onUpdate = ForeignKeyAction.Cascade)
    def supidx = index("SUP_IDX",supid,unique = false)
    def nidx = index("NM_IDX",name,unique = true)
  }
  val coffees = TableQuery[Coffees]

}

src/main/scala/access/DAOs.scala:

package com.bayakala.learn.slick301.access
import com.bayakala.learn.slick301.config
import com.bayakala.learn.slick301.config.DBConfig
import com.bayakala.learn.slick301.model.TableDefs
trait DAOs { dbconf: DBConfig =>
  import jdbcDriver.api._
  //注入依赖
  val tables = new TableDefs(dbconf.jdbcDriver)
  import tables._
  //suppliers queries
  val createSupplierTable = suppliers.schema.create
  val allSuppliers = suppliers.result
  def insertSupplier(id:Long,name:String,address:Option[String],website:Option[String])
  = suppliers += Supplier(id,name,address,website)
  def insertSupbyName(n: String) = suppliers.map(_.name) += n
  //coffees queries
  val createCoffeeTable = coffees.schema.create
  val allCoffees = coffees.result
  def insertCoffee(c: (Long,String,Long,Double,Int)) =
    coffees += Coffee(id=c._1, name=c._2,supid=c._3,price=c._4,sales=c._5)

}

src/main/scala/main/Main.scala:

package com.bayakala.learn.slick301.main
import com.bayakala.learn.slick301.config.DBConfig
import com.bayakala.learn.slick301.access.DAOs

import scala.concurrent.Await
import scala.util.{Failure, Success}
import scala.concurrent.duration._
import scala.concurrent.ExecutionContext.Implicits.global
import slick.backend.DatabaseConfig
import slick.driver.JdbcProfile

object Main {

  object Actions extends DAOs with DBConfig {
    import slick.SlickException
    import slick.util.ClassLoaderUtil
    import scala.util.control.NonFatal
    import com.typesafe.config.ConfigFactory

    def getDbConfig: String =
        ConfigFactory.load().getString("app.dbconfig")

    def getDbDriver(path: String): JdbcProfile = {
      val config = ConfigFactory.load()
      val n = config.getString((if (path.isEmpty) "" else path + ".") + "driver")
      val untypedP = try {
        if (n.endsWith("$")) ClassLoaderUtil.defaultClassLoader.loadClass(n).getField("MODULE$").get(null)
        else ClassLoaderUtil.defaultClassLoader.loadClass(n).newInstance()
      } catch {
        case NonFatal(ex) =>
          throw new SlickException(s"""Error getting instance of Slick driver "$n"""", ex)
      }
      untypedP.asInstanceOf[JdbcProfile]
    }

    override lazy val jdbcDriver: JdbcProfile = getDbDriver(getDbConfig)
    val dbConf: DatabaseConfig[JdbcProfile] = DatabaseConfig.forConfig(getDbConfig)
    override val db = dbConf.db
  }
  import Actions._


  def main(args: Array[String]) = {

    val res = db.run(createSupplierTable).andThen {
      case Success(_) => println("supplier table created")
      case Failure(_) => println("unable to create supplier table")
    }
    Await.ready(res, 3 seconds)

    val res2 = db.run(insertSupbyName("Acme Coffee Co."))
    Await.ready(res2, 3 seconds)

    Await.ready(db.run(allSuppliers), 10 seconds).foreach(println)

    val res10 = db.run(createCoffeeTable).andThen {
      case Success(_) => println("coffee table created")
      case Failure(_) => println("unable to create coffee table")
    }
    Await.ready(res10, 3 seconds)

    val res11 = db.run(insertCoffee((101,"Columbia",1,158.0,0)))
    Await.ready(res11, 3 seconds)

    Await.ready(db.run(allCoffees), 10 seconds).foreach(println)

  }