Rick Lin created SPARK-17775:
--------------------------------
Summary: pyspark: take(num) failed, but collect() worked for big
dataset
Key: SPARK-17775
URL: https://issues.apache.org/jira/browse/SPARK-17775
Project: Spark
Issue Type: Bug
Environment: Spark:1.6.1
Python 2.7.12 :: Anaconda 4.1.1 (64-bit)
Windows 7
One machine
Reporter: Rick Lin
Hi, all:
I ran one dataset with 39,501 data drew from the table of PostgreSQL DB in
pyspark.
The code was as:
cur1.execute("select id from users")
users = cur1.fetchall()
users_rdd = sc.parallelize(users)
users_rdd.take(1)
where the error message was as:
Py4JJavaError: An error occurred while calling
z:org.apache.spark.api.python.PythonRDD.runJob.
: org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in
stage 0.0 failed 1 times, most recent failure: Lost task 0.0 in stage 0.0 (TID
0, localhost): java.net.SocketException: Connection reset by peer: socket write
error
However, when i changed from take(1) to collect(), it worked, as:
[[25],
[1439],
...
]
When I ran the same code for a small dataset, here take(1) and collect() worked.
I don't know why this happened and how to fix this problem for big dataset?
Could you help me to deal with this problem?
Thanks
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
