Question

我想写一个Spark Streamin App，它使用随机整数的流来计算它们。这是我写的Spark应用程序：

from pyspark import SparkContext
from pyspark.streaming import StreamingContext

sc = SparkContext("local[2]", "IntegerCount") # 2 threads, app name
ssc = StreamingContext(sc, 1) # sc, time interval for batch update.

nums = ssc.socketTextStream("localhost", 8000) # stream data from TCP; source, port

# create key,value pairs
tests = nums.map(lambda num: (int(num), 1))

# Count each integer in each batch
intCounts = tests.reduceByKey(lambda x, y: x + y)

# Print
intCounts.pprint()

ssc.start()             # Start the computation
ssc.awaitTermination()  # Wait for the computation to terminate

我正在使用该Server.py：

向端口8000提供随机数

import socket
from random import randint

host = 'localhost'
port = 8000
address = (host, port)

server_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
server_socket.bind(address)
server_socket.listen(5)


print "Listening for client . . ."
conn, address = server_socket.accept()
print "Connected to client at ", address
#pick a large output buffer size because i dont necessarily know how big the incoming packet is
while True:
    output = str(randint(0, 10))
    conn.send(output)

当我运行Server.py和我的Spark App时，连接成功建立。但是我看到一个空输出：

Using Spark's default log4j profile: org/apache/spark/log4j-defaults.properties
Setting default log level to "WARN".
To adjust logging level use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).
-------------------------------------------
Time: 2017-07-16 22:36:11
-------------------------------------------

-------------------------------------------
Time: 2017-07-16 22:36:12
-------------------------------------------

我不知道问题是什么，请帮我理解发生了什么？

Answer 1

解决了，我用“\ n”发送字符串并且它有效。

import socket
from random import randint

host = 'localhost'
port = 8000
address = (host, port)

server_socket = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
server_socket.bind(address)
server_socket.listen(5)


print "Listening for client . . ."
conn, address = server_socket.accept()
print "Connected to client at ", address
#pick a large output buffer size because i dont necessarily know how big the incoming packet is
while True:
    output = str(randint(0, 10)) + "\n"  ### THAT IS THE FIX.
    conn.send(output)

Spark Streaming App无法在端口上接收字符串

1 个答案: