Ameen Tayyebi created SPARK-20588:
-------------------------------------
Summary: from_utc_timestamp causes bottleneck
Key: SPARK-20588
URL: https://issues.apache.org/jira/browse/SPARK-20588
Project: Spark
Issue Type: Bug
Components: SQL
Affects Versions: 2.0.2
Environment: AWS EMR AMI 5.2.1
Reporter: Ameen Tayyebi
We have a SQL query that makes use of the from_utc_timestamp function like so:
from_utc_timestamp(itemSigningTime,'America/Los_Angeles')
This causes a major bottleneck. Our exact call is:
date_add(from_utc_timestamp(itemSigningTime,'America/Los_Angeles'), 1)
Switching from the above to date_add(itemSigningTime, 1) reduces the job
running time from 40 minutes to 9.
When from_utc_timestamp function is used, several threads in the executors are
in the BLOCKED state, on this call stack:
"Executor task launch worker-63" #261 daemon prio=5 os_prio=0
tid=0x00007f848472e000 nid=0x4294 waiting for monitor entry [0x00007f501981c000]
java.lang.Thread.State: BLOCKED (on object monitor)
at java.util.TimeZone.getTimeZone(TimeZone.java:516)
- waiting to lock <0x00007f5216c2aa58> (a java.lang.Class for
java.util.TimeZone)
at
org.apache.spark.sql.catalyst.util.DateTimeUtils$.stringToTimestamp(DateTimeUtils.scala:356)
at
org.apache.spark.sql.catalyst.util.DateTimeUtils.stringToTimestamp(DateTimeUtils.scala)
at
org.apache.spark.sql.catalyst.expressions.GeneratedClass$GeneratedIterator.agg_doAggregateWithKeys$(Unknown
Source)
at
org.apache.spark.sql.catalyst.expressions.GeneratedClass$GeneratedIterator.processNext(Unknown
Source)
at
org.apache.spark.sql.execution.BufferedRowIterator.hasNext(BufferedRowIterator.java:43)
at
org.apache.spark.sql.execution.WholeStageCodegenExec$$anonfun$8$$anon$1.hasNext(WholeStageCodegenExec.scala:370)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:408)
at
org.apache.spark.shuffle.sort.UnsafeShuffleWriter.write(UnsafeShuffleWriter.java:161)
at
org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:79)
at
org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:47)
at org.apache.spark.scheduler.Task.run(Task.scala:86)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:274)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)
Can we cache the locale's once per JVM so that we don't do this for every
record?
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
