我使用光滑的codegen自动生成了scala代码。我看到一些表Rows是作为HLists实现的。 (但这些是光滑的HList,而不是正常的无形HList)
现在我希望通过光滑查询将HList中的特定元素作为Row返回。
我用Google搜索并找到了这个帖子
Getting elements from an HList
但这对于光滑的HList不起作用。它适用于Shapeless HList
我也尝试过apply方法
val x:Long = slickHList(2)
但是这不能编译,因为类型Any不符合Long的执行类型。我不想做.asInstanceOf
是否有类型安全的方法可以访问光滑的HList元素?
编辑:根据下面的输入,我编写了以下代码
package com.abhi
object SlickAndShapeless {
import slick.collection.heterogeneous.{HCons, HList, HNil}
import slick.collection.heterogeneous.syntax.HNil
type MyRow = HCons[Long, HCons[String, HNil]]
val row : MyRow = 1L :: "foo" :: HNil
import HListExtensions._
val hlist = row.asShapeless
val container = new Container(hlist)
val item = container.get(1)
}
class Container[L <: shapeless.HList](list: L) {
import shapeless._
import nat._
import ops.hlist._
def get(n: Nat)(implicit at: At[L, n.N]): at.Out = list[n.N]
}
object HListExtensions {
import slick.collection.heterogeneous.{HNil => SHNil, HList => SHList, HCons}
import shapeless.{::, HList, HNil}
implicit class HListShapelessSlick(val list: HList) extends AnyVal {
def asSlick : SHList = list match {
case HNil => SHNil
case head :: tail => head :: tail.asSlick
}
}
implicit class HListSlickShapeless(val list: SHList) extends AnyVal {
def asShapeless : HList = list match {
case SHNil => HNil
case HCons(head, tail) => head :: tail.asShapeless
}
}
}
上面代码的问题在于,从item
获得的val item = container.get(1)
类型为at.Out
,而不是我期待的Long
。
build.sbt
libraryDependencies ++= Seq(
"com.typesafe.slick" % "slick_2.12" % "3.2.1",
"com.chuusai" % "shapeless_2.12" % "2.3.2"
)
我也看到两个编译器错误
Error:(19, 35) Implicit not found: shapeless.Ops.At[shapeless.HList, shapeless.Succ[shapeless._0]]. You requested to access an element at the position shapeless.Succ[shapeless._0], but the HList shapeless.HList is too short.
val item : Long = container.get(1)
Error:(19, 35) not enough arguments for method get: (implicit at: shapeless.ops.hlist.At[shapeless.HList,shapeless.Succ[shapeless._0]])at.Out.
Unspecified value parameter at.
val item : Long = container.get(1)
答案 0 :(得分:3)
可以创建扩展方法:
nppSetStream(cudaStreamPerThread)
示例:
object HListExtensions {
import slick.collection.heterogeneous.{HNil => SHNil, HList => SHList, HCons}
import shapeless.{ ::, HList, HNil }
implicit class HListShapelessSlick(val list:HList) extends AnyVal {
def asSlick:SHList = list match {
case HNil => SHNil
case head :: tail => head :: tail.asSlick
}
}
implicit class HListSlickShapeless(val list:SHList) extends AnyVal {
def asShapeless:HList = list match {
case SHNil => HNil
case HCons(head, tail) => head :: tail.asShapeless
}
}
}
我希望这会有所帮助。
编辑:这是类型级解决方案。
scala>import HListExtensions._
import HListExtensions._
scala> val x1:HList = 1 :: 2 :: HNil
x1: slick.collection.heterogeneous.HList = 1 :: 2 :: HNil
scala> x1.asShapeless
res1: shapeless.HList = 1 :: 2 :: HNil
scala> x1.asShapeless.asSlick
res2: slick.collection.heterogeneous.HList = 1 :: 2 :: HNil
示例:
object HListsConvertersTypeLevel {
import shapeless.{::}
sealed trait HConv[From <: heterogeneous.HList, To <: shapeless.HList] {
def convert(list: From): To
}
implicit def buildHConvNil: HConv[heterogeneous.HNil.type, shapeless.HNil] =
new HConv[heterogeneous.HNil.type, shapeless.HNil] {
override def convert(list: heterogeneous.HNil.type): shapeless.HNil = shapeless.HNil
}
implicit def buildHConv[H, T <: heterogeneous.HList, T2 <: shapeless.HList](
implicit conv: HConv[T, T2]): HConv[HCons[H, T], ::[H, T2]] = new HConv[HCons[H, T], ::[H, T2]] {
override def convert(list: HCons[H, T]): ::[H, T2] = {
list.head :: conv.convert(list.tail)
}
}
def toShapeless[A <: heterogeneous.HList, B <: shapeless.HList](list: A)(implicit conv: HConv[A, B]): B = conv.convert(list)
}