- 對於大批量數據,查詢es時,需要帶條件去查詢,否則一下查出所有數據數據量會很大
- es查詢需要編寫json格式的DSL查詢語句,對於複雜查詢,DSL編寫起來也分很複雜,所以我們這裏使用sparksql,通過編寫sql語句,spark將sql語句自動轉化爲DSL語句來查詢es。
關於es對spark的支持可參見文檔,官網 https://www.elastic.co/guide/en/elasticsearch/hadoop/current/spark.html
廢話不多說,上代碼
case class people(phone:String)
val sparkConf = new SparkConf().setAppName("application-name").setMaster("local[4]")
val sc = new SparkContext(sparkConf)
val sql = new SQLContext(sc)
import sql.implicits._
//配置
val options = Map(
"pushdown" -> "true",
"es.nodes" -> "192.45.15.15", //ip
"es.port" -> "9200") //端口
val esDF = sql.read.format("org.elasticsearch.spark.sql")
.options(options)
.load("spark_202003/alldata") //索引
// esDF.printSchema() //可打印schema信息
val table = esDF.createOrReplaceTempView("spark_202003") //創建臨時表用於查詢
val sqlString=“select * from spark_202003 where code=‘1’ ”
val resultDF = sql.sql(sqlString)
var resultDS = resultDF.map(x => people(x(0).toString))) //
這裏是轉換成了dataset,之後可對dataset 做其他操作
打印下查詢的內容
resultDS.collect().toList.toString()