1、broadcast广播
在Spark中，broadcast是一种优化技术，它可以将一个只读变量缓存到每个节点上，以便在执行任务时使用。这样可以避免在每个任务中重复传输数据。
2、构建缓存
import org.apache.spark.sql.SparkSession
import org.apache.spark.broadcast.Broadcast
import com.alibaba.fastjson.JSONObject// 定义全局缓存单例对象
object GlobalCache extends Serializable {// 广播变量，用于存储缓存数据private var cacheData: Broadcast[collection.mutable.Map[String, JSONObject]] = _// 设置 SparkSession 和广播变量def setSparkSession(spark: SparkSession): Unit = {cacheData = spark.sparkContext.broadcast(collection.mutable.Map.empty[String, JSONObject])}// 按订单ID和用户ID缓存JSONObject对象def cacheJSONObject(orderId: String, userId: String, jsonObject: JSONObject): Unit = {// 获取广播变量的值并进行修改val data = cacheData.valuedata.synchronized {data.put(generateKey(orderId, userId), jsonObject)}}// 根据订单ID和用户ID删除缓存的JSONObject对象def removeJSONObject(orderId: String, userId: String): Unit = {// 获取广播变量的值并进行修改val data = cacheData.valuedata.synchronized {data.remove(generateKey(orderId, userId))}}// 根据订单ID和用户ID获取缓存的JSONObject对象def getJSONObjet(orderId: String, userId: String): JSONObject = {// 获取广播变量的值并进行访问val data = cacheData.valuedata.synchronized {data.get(generateKey(orderId, userId)).orNull}}// 生成缓存键，使用订单ID和用户ID拼接private def generateKey(orderId: String, userId: String): String = s"$orderId|$userId"
}
3、缓存测试
import org.apache.spark.sql.SparkSession
import org.apache.spark.broadcast.Broadcast
import com.alibaba.fastjson.JSONObject
import org.apache.log4j.{Level, Logger}object CacheTest {Logger.getLogger("org").setLevel(Level.ERROR)Logger.getRootLogger().setLevel(Level.ERROR) // 设置日志级别def addItem(orderId:String, userId:String, name:String): Unit = {val jsonObject = new JSONObject()jsonObject.put("name", name)// 缓存JSONObject对象GlobalCache.cacheJSONObject(orderId, userId, jsonObject)}def getCache(orderId: String, userId: String): JSONObject = {// 获取缓存的JSONObject对象GlobalCache.getJSONObjet(orderId, userId)}def delItem(orderId:String, userId:String): Unit = {// 删除缓存的JSONObject对象GlobalCache.removeJSONObject(orderId, userId)}def getSparkSession(appName: String, localType: Int): SparkSession = {val builder: SparkSession.Builder = SparkSession.builder().appName(appName)if (localType == 1) {builder.master("local[8]") // 本地模式，启用8个核心}val spark = builder.getOrCreate() // 获取或创建一个新的SparkSessionspark.sparkContext.setLogLevel("ERROR") // Spark设置日志级别spark}def main(args: Array[String]): Unit = {println("Start CacheTest")val spark: SparkSession = getSparkSession("CacheTest", 1)GlobalCache.setSparkSession(spark) // 构造全局缓存addItem("001", "456", "苹果") // 添加元素addItem("002", "789", "香蕉") // 添加元素var cachedObject = getCache("001", "456")println(s"Cached Object: $cachedObject")delItem("001", "456") // 删除元素cachedObject = getCache("001", "456")println(s"Cached Object: $cachedObject")spark.stop()}
}
4、控制台输出
Start CacheTest
Using Spark's default log4j profile: org/apache/spark/log4j-defaults.properties
Cached Object: {"name":"苹果"}
Cached Object: nullProcess finished with exit code 0
