search（6）- elastic4s-CRUD

    如果我们把ES作为某种数据库来使用的话，必须熟练掌握ES的CRUD操作。在这之前先更正一下上篇中关于检查索引是否存在的方法：elastic4s的具体调用如下：

  //删除索引
  val rspExists = client.execute(indexExists("company")).await
  if (rspExists.result.exists)
     client.execute(deleteIndex("company")).await

在下面我们还会示范如何检查一条记录（document）是否存在的方法。

先示范新建一条记录。一般来讲数据库表都有个唯一字段，最好用ES里的id来代表，否则ES会自动产生一个唯一id，那么随机读取get时就会很不方便。如果新插入的记录id已经在表里存在，ES会替换新的内容，不会产生异常。可以在elastic4s里使用createOnly(true)来强制产生重复id异常：

import com.sksamuel.elastic4s.{ElasticClient, ElasticProperties}
import com.sksamuel.elastic4s.akka._
import akka.actor._

import scala.concurrent.ExecutionContext.Implicits.global
object Lesson05 extends App {
  import com.sksamuel.elastic4s.ElasticDsl._

  private implicit lazy val system: ActorSystem = ActorSystem()
  val akkaClient = AkkaHttpClient(AkkaHttpClientSettings(List("130.1.1.234:9200")))
  val client = ElasticClient(akkaClient)

  val publisher = indexInto("company").id("c00001")
    .fields(
      "code" -> "c00001",
      "name" -> "人民出版社",
      "biztype" -> "出版社",
      "addr" -> Map(
        "district" -> "北京市东城区",
        "address" -> "朝阳门内大街166号"
      ),
      "regdate" -> "1963-02-18",
      "contact" -> "65122634@163.com"
    ).createOnly(true)

  val pubExists = client.execute(exists("company","c00001")).await
  if (pubExists.isSuccess) {
    val createPub = client.execute(publisher).await

    if (createPub.isSuccess) {
      val pub = client.execute(get("company", "c00001").fetchSourceContext(true)).await
      println(s"${pub.result.sourceAsMap}")
    } else println(s"${createPub.error.reason}")
  } else println(s"${pubExists.error.reason}")

  val dstributor = indexInto("company").id("c00002")
    .fields(
      "code" -> "c00002",
      "name" -> "新华文轩出版传媒股份有限公司",
      "biztype" -> "出版发行",
      "addr" -> Map(
        "district" -> "四川省成都市锦江区",
        "address" -> "金石路239号4栋1层1号"
      ),
      "regdate" -> "2005-03-09",
      "contact" -> "52635286@qq.com"
    ).createOnly(true)

  val grpExists = client.execute(exists("company","c00002")).await
  if (grpExists.isSuccess) {
    val createGroup = client.execute(dstributor).await

    if (createGroup.isSuccess) {
      val dstr = client.execute(get("company", "c00002").fetchSourceContext(true)).await
      println(s"${dstr.result.sourceAsMap}")
    } else println(s"${createGroup.error.reason}")
  } else println(s"${grpExists.error.reason}")

  val mget = client.execute(multiget(
    get("company","c00001"),
    get("company","c00002")
  )).await
  if(mget.isSuccess)
    mget.result.items.foreach(i => println(s"${i.sourceAsMap}"))
  else println(s"${mget.error.reason}")

  scala.io.StdIn.readLine()
  system.terminate()
  client.close()
}

上面示范了不同类型字段的填写方式，特别是nested字段如addr。每插入一条新记录就用get进行一次验证，输出显示：

HashMap(name -> 人民出版社, regdate -> 1963-02-18, contact -> 65122634@163.com, code -> c00001, addr -> Map(district -> 北京市东城区, address -> 朝阳门内大街166号), biztype -> 出版社)
HashMap(name -> 新华文轩出版传媒股份有限公司, regdate -> 2005-03-09, contact -> 52635286@qq.com, code -> c00002, addr -> Map(district -> 四川省成都市锦江区, address -> 金石路239号4栋1层1号), biztype -> 出版发行)
HashMap(name -> 人民出版社, regdate -> 1963-02-18, contact -> 65122634@163.com, code -> c00001, addr -> Map(district -> 北京市东城区, address -> 朝阳门内大街166号), biztype -> 出版社)
HashMap(name -> 新华文轩出版传媒股份有限公司, regdate -> 2005-03-09, contact -> 52635286@qq.com, code -> c00002, addr -> Map(district -> 四川省成都市锦江区, address -> 金石路239号4栋1层1号), biztype -> 出版发行)

上面提到过，如果我们想把ES当作普通的数据库来使用的话，还是要备齐了CRUD功能。具体操作按照关系数据库方式围绕着唯一键id进行。比如，我们可以用id来检查记录是否已经存在：

  val pubExists = client.execute(exists("company","c00001")).await
  if (pubExists.isSuccess) {...}

上面我们示范了针对索引的create,read操作。下面讨论一下update：update 可分单笔或批次两类，分别为：updateById, updateByQuery，很明显：updateByQuery是以query作为目标筛选条件的成批update操作。与上面的create操作一样，我们还是需要考虑唯一键id，这个可以在updateById操作里处理：当目标id存在时，用update请求里的字段值更新对应的字段。如目标id不存在的话就把update请求里的字段值当作新记录内容插入：

import com.sksamuel.elastic4s.http.JavaClient
import com.sksamuel.elastic4s.requests.common.RefreshPolicy
import com.sksamuel.elastic4s.{ElasticClient, ElasticProperties}
import scala.concurrent.ExecutionContext.Implicits.global
object Lesson06 extends App {
  import com.sksamuel.elastic4s.ElasticDsl._

  val esjava = JavaClient(ElasticProperties("http://localhost:9200"))
  val client = ElasticClient(esjava)

  val doc1 = updateById("company","t00001")
      .docAsUpsert(
        Map(
          "code" -> "t00001",
          "name" -> "test company1"
        )
      )

  val doc2 = updateById("company","t00002")
    .docAsUpsert(
      Map(
        "code" -> "t00002",
        "name" -> "test company2"
      )
    )

  val doc3 = updateById("company","t00003")
    .docAsUpsert(
      Map(
        "code" -> "t00003",
        "name" -> "test company3"
      )
    )

  val updateAll = for {
    _ <- client.execute(doc1)
    _ <- client.execute(doc2)
    _ <- client.execute(doc3)
  } yield()
  updateAll.await

  val getResults = client.execute(multiget(
     get("company","t00001").fetchSourceInclude("code","name"),
     get("company","t00002").fetchSourceInclude("code","name"),
     get("company","t00003").fetchSourceInclude("code","name")
   )
  ).await

  getResults.result.items.foreach(i => println(i.sourceAsMap))

  client.close()
}

成批更新比较麻烦，因为通常每条记录的更新都可能涉及到当前记录的字段值，或作为判断条件，或为更新值，我们需要使用并处理当前记录中某些字段。这就需要在数据层面运行某些计算方法，可以用脚本语言来实现这样的功能，如下：

  import com.sksamuel.elastic4s.requests.script.Script
  val script = "ctx._source.fullname = ctx._source.code+' '+ctx._source.name"
  val updateByQ = updateIn("company")
    .query(matchQuery("name","test"))
    .script(Script(script,Some("painless")))

  val qupResult = client.execute(updateByQ).await

  val getResults = client.execute(multiget(
     get("company","t00001"),
     get("company","t00002"),
     get("company","t00003")
   )
  ).await

  getResults.result.items.foreach(i => println(i.sourceAsMap))

与update一样，delete也分单个或成批删除模式。delete by Id 示例如下：

  (for {
    _ <- client.execute(delete("t00001").from("company"))
    _ <- client.execute(deleteByQuery("company", "t00002"))
  } yield()).await

delete by Query 用法如下：

  import com.sksamuel.elastic4s.Index._
  client.execute(
    deleteByQuery(toIndex("company"),
      termQuery("code","t00003"))
  ).await