将 Spark Streaming 数据发送回客户端

Question

I am new to Apache Spark Streaming. I am developing a spark streaming application to find the shortest path and sending the path again back to the client. I have written code for taking data and processing it with a function but I have an issue how can I send my results again back to the client Here is my code:

import networkx as nx
from pyspark import SparkConf,SparkContext
from pyspark.streaming import StreamingContext
TCP_IP = "127.0.0.1"
TCP_PORT = 5000

# Creating a Spark Configuration
conf=SparkConf()
conf.setAppName('ShortestPathApp')

sc= SparkContext(conf)
ssc= StreamingContext(sc,2)

def shortestPath(line):
    # get the values from rdd
    vehicleId = line[0]
    source = line[1]
    destination = line[2]
    deadline = line[3]

    # find shortest path
    shortest = nx.dijkstra_path(G, source, destination)



# receive from Socket
dataStream =ssc.socketTextStream(TCP_IP,TCP_PORT)
vehicle_data = dataStream.map(lambda line: line.split(" "))
vehicle_data.foreachRDD(lambda rdd: rdd.foreach(shortestPath))
ssc.start()
ssc.awaitTermination()

How can I send the data back to the client

Answer 1

Use StreamingContext to push output data back to destination as a stream. you can create it as below. Add method to get singleton SparkSession instance

# Lazily instantiated global instance of SparkSession
def getSparkSessionInstance(sparkConf):
    if ("sparkSessionSingletonInstance" not in globals()):
        globals()["sparkSessionSingletonInstance"] = SparkSession \
            .builder \
            .config(conf=sparkConf) \
            .getOrCreate()
    return globals()["sparkSessionSingletonInstance"]

sparkSess = getSparkSessionInstance(rdd.context.getConf()) 
vehicle_data_df = sparkSess.createDataFrame(vehicle_data)
vehicle_data_df.writeStream\
      .format("socket")\
      .option("host",TCP_OUTPUT_IP)    //Output socket IP address 
      .option("port",TCP_OUTPUT_PORT)  //Output socket port 
      .outputMode('append')\
      .start()\
      .awaitTermination()