Slick (Scala language-integrated connection kit)是scala的一个FRM(Functional Relational Mapper),即函数式的关系数据库编程工具库。Slick的主要目的是使关系数据库能更容易、更自然的融入函数式编程模式,它可以使使用者像对待scala集合一样来处理关系数据库表。也就是说可以用scala集合的那些丰富的操作函数来处理库表数据。Slick把数据库编程融入到scala编程中,编程人员可以不需要编写SQL代码。我把Slick官方网站上Slick3.1.1文档的Slick介绍章节中的一些描述和例子拿过来帮助介绍Slick的功能。下面是Slick数据库和类对象关系对应的一个例子:
import slick.driver.H2Driver.api._
object slickIntro {
case class Coffee(id: Int,
name: String,
supID: Int = 0,
price: Double ,
sales: Int = 0,
total: Int = 0)
class Coffees(tag: Tag) extends Table[Coffee](tag, "COFFEES") {
def id = column[Int]("COF_ID", O.PrimaryKey, O.AutoInc)
def name = column[String]("COF_NAME")
def supID = column[Int]("SUP_ID")
def price = column[Double]("PRICE")
def sales = column[Int]("SALES", O.Default(0))
def total = column[Int]("TOTAL", O.Default(0))
def * = (id, name, supID, price, sales, total) <> (Coffee.tupled, Coffee.unapply)
}
val coffees = TableQuery[Coffees]
//> coffees : slick.lifted.TableQuery[worksheets.slickIntro.Coffees] = Rep(TableExpansion)
}
我们把数据库中的COFFEES表与Coffees类做了对应,包括字段、索引、默认值、返回结果集字段等。现在这个coffees就是scala里的一个对象,但它代表了数据库表。现在我们可以用scala语言来编写数据存取程序了:
val limit = 10.0 //> limit : Double = 10.0
// // 写Query时就像下面这样:
( for( c <- coffees; if c.price < limit ) yield c.name ).result
//> res0: slick.driver.H2Driver.StreamingDriverAction[Seq[String],String,slick.dbio.Effect.Read] = slick.driver.JdbcActionComponent$QueryActionExtensionMethodsImpl$[email protected]
// 相当于 SQL: select COF_NAME from COFFEES where PRICE < 10.0
或者下面这些不同的Query:
// 返回"name"字段的Query
// 相当于 SQL: select NAME from COFFEES
coffees.map(_.name)
//> res1: slick.lifted.Query[slick.lifted.Rep[String],String,Seq] = Rep(Bind)
// 选择 price < 10.0 的所有记录Query
// 相当于 SQL: select * from COFFEES where PRICE < 10.0
coffees.filter(_.price < 10.0)
//> res2: slick.lifted.Query[worksheets.slickIntro.Coffees,worksheets.slickIntro.Coffees#TableElementType,Seq] = Rep(Filter @1946988038)
我们可以这样表述:coffees.map(_.name) >>> coffees.map{row=>row.name}, coffees.filter(_.price<10.0) >>> coffees.filter{row=>row.price<10.0),都是函数式集合操作语法。
Slick把Query编写与scala语言集成,这使编程人员可以用熟悉惯用的scala来表述SQL Query,直接的好处是scalac在编译时就能够发现Query错误:
//coffees.map(_.prices) //编译错误:value prices is not a member of worksheets.slickIntro.Coffees
当然,嵌入scala的Query还可以获得运行效率的提升,因为在编译时可以进行前期优化。
最新版本的Slick最大的特点是采用了Functional I/O技术,从而实现了安全的多线程无阻碍I/O操作。再就是实现了Query的函数组合(functional composition),使Query编程更贴近函数式编程模式。通过函数组合实现代码重复利用,提高编程工作效率。具体实现方式是利用freemonad(DBIOAction类型就是个freemonad)的延迟运算模式,将DBIOAction的编程和实际运算分离,在DBIOAction编程过程中不会产生副作用(side-effect),从而实现纯代码的函数组合。我们来看看Query函数组合和DBIOAction运算示范:
import scala.concurrent.ExecutionContext.Implicits.global
val qDelete = coffees.filter(_.price > 0.0).delete
//> qDelete : slick.driver.H2Driver.DriverAction[Int,slick.dbio.NoStream,slick.dbio.Effect.Write] ...
val qAdd1 = (coffees returning coffees.map(_.id)) += Coffee(name="Columbia",price=128.0)
//> qAdd1 : slick.profile.FixedSqlAction[Int,slick.dbio.NoStream,slick.dbio.Effect.Write]...
val qAdd2 = (coffees returning coffees.map(_.id)) += Coffee(name="Blue Mountain",price=828.0)
//> qAdd2 : slick.profile.FixedSqlAction[Int,slick.dbio.NoStream,slick.dbio.Effect.Write]...
def getNameAndPrice(n: Int) = coffees.filter(_.id === n)
.map(r => (r.name,r.price)).result.head
//> getNameAndPrice: (n: Int)slick.profile.SqlAction[(String, Double),slick.dbio.NoStream,slick.dbio.Effect.Read]
val actions = for {
_ <- coffees.schema.create
_ <- qDelete
c1 <- qAdd1
c2 <- qAdd2
(n1,p1) <- getNameAndPrice(c1)
(n2,p2) <- getNameAndPrice(c2)
} yield (n1,p1,n2,p2)
//> actions : slick.dbio.DBIOAction[(String, Double, String, Double),..
我们可以放心的来组合这个actions,不用担心有任何副作用。actions的类型是:DBAction[String,Double,String,Double]。我们必须用Database.Run来真正开始运算,产生副作用:
import java.sql.SQLException
import scala.concurrent.Await
import scala.concurrent.duration._
val db = Database.forURL("jdbc:h2:mem:demo", driver="org.h2.Driver")
//> db : slick.driver.H2Driver.backend.DatabaseDef = [email protected]
Await.result(
db.run(actions.transactionally).map { res =>
println(s"Add coffee: ${res._1},${res._2} and ${res._3},${res._4}")
}.recover {
case e: SQLException => println("Caught exception: " + e.getMessage)
}, Duration.Inf) //> Add coffee: Columbia,128.0 and Blue Mountain,828.0
在特殊的情况下我们也可以引用纯SQL语句:Slick提供了Plain SQL API, 如下:
val limit = 10.0 sql"select COF_NAME from COFFEES where PRICE < $limit".as[String] // 用$来绑定变量: // select COF_NAME from COFFEES where PRICE < ?
下面是这篇讨论的示范代码:
package worksheets import slick.driver.H2Driver.api._ object slickIntro { case class Coffee(id: Int = 0, name: String, supID: Int = 0, price: Double, sales: Int = 0, total: Int = 0) class Coffees(tag: Tag) extends Table[Coffee](tag, "COFFEES") { def id = column[Int]("COF_ID", O.PrimaryKey, O.AutoInc) def name = column[String]("COF_NAME") def supID = column[Int]("SUP_ID") def price = column[Double]("PRICE") def sales = column[Int]("SALES", O.Default(0)) def total = column[Int]("TOTAL", O.Default(0)) def * = (id, name, supID, price, sales, total) <> (Coffee.tupled, Coffee.unapply) } val coffees = TableQuery[Coffees] val limit = 10.0 // // 写Query时就像下面这样: ( for( c <- coffees; if c.price < limit ) yield c.name ).result // 相当于 SQL: select COF_NAME from COFFEES where PRICE < 10.0 // 返回"name"字段的Query // 相当于 SQL: select NAME from COFFEES coffees.map(_.name) // 选择 price < 10.0 的所有记录Query // 相当于 SQL: select * from COFFEES where PRICE < 10.0 coffees.filter(_.price < 10.0) //coffees.map(_.prices) //编译错误:value prices is not a member of worksheets.slickIntro.Coffees import scala.concurrent.ExecutionContext.Implicits.global val qDelete = coffees.filter(_.price > 0.0).delete val qAdd1 = (coffees returning coffees.map(_.id)) += Coffee(name="Columbia",price=128.0) val qAdd2 = (coffees returning coffees.map(_.id)) += Coffee(name="Blue Mountain",price=828.0) def getNameAndPrice(n: Int) = coffees.filter(_.id === n) .map(r => (r.name,r.price)).result.head val actions = for { _ <- coffees.schema.create _ <- qDelete c1 <- qAdd1 c2 <- qAdd2 (n1,p1) <- getNameAndPrice(c1) (n2,p2) <- getNameAndPrice(c2) } yield (n1,p1,n2,p2) import java.sql.SQLException import scala.concurrent.Await import scala.concurrent.duration._ val db = Database.forURL("jdbc:h2:mem:demo", driver="org.h2.Driver") Await.result( db.run(actions.transactionally).map { res => println(s"Add coffee: ${res._1},${res._2} and ${res._3},${res._4}") }.recover { case e: SQLException => println("Caught exception: " + e.getMessage) }, Duration.Inf) }
原创文章,作者:ItWorker,如若转载,请注明出处:https://blog.ytso.com/12886.html