[HIVEMALL-219] NullPointerException in pLSA model

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.5.0
• Fix Version/s: 0.5.2
• Assignee: Makoto Yui
• Reporter: jsocsmao
• Created: 17/Sep/18 19:12
• Updated: 18/Sep/18 10:50
• Resolved: 18/Sep/18 10:47

Description

 

I was not able to have pLSA model working on relatively small production workload (5000 docs) with number of reducers > 1.

It worked with 1000 documents in a single reducer.

 

I tried Tez or map-reduce engines:

hive> create table docres as
    > with word_counts as (
    >   select docid, feature(word, count(word)) as f
    >   from docs t1 lateral view explode(tokenize(doc, true)) t2 as word
    >   where not is_stopword(word)
    >   group by docid, word
    > ),
    > input as (
    >   select docid, collect_list(f) as features
    >   from word_counts
    >   group by docid)
    > select  train_plsa(features, '-topics 20 -iter 10 -s 500 -delta 1 -alpha 500 -eps 1') as (label, word, prob)
    > from  input;
Total jobs = 1
Launching Job 1 out of 1
Number of reduce tasks not specified. Defaulting to jobconf value of: 4
In order to change the average load for a reducer (in bytes):
  set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
  set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
  set mapreduce.job.reduces=<number>

Kill Command = /usr/hdp/2.6.1.10-4/hadoop/bin/hadoop job  -kill job_1536748924580_2024
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 4
2018-09-17 19:05:38,300 Stage-1 map = 0%,  reduce = 0%
2018-09-17 19:05:52,225 Stage-1 map = 35%,  reduce = 0%, Cumulative CPU 10.56 sec
2018-09-17 19:05:55,428 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 14.69 sec
2018-09-17 19:06:01,903 Stage-1 map = 100%,  reduce = 25%, Cumulative CPU 14.69 sec
2018-09-17 19:06:02,963 Stage-1 map = 100%,  reduce = 0%, Cumulative CPU 14.69 sec
2018-09-17 19:06:09,433 Stage-1 map = 100%,  reduce = 17%, Cumulative CPU 25.22 sec
2018-09-17 19:06:12,803 Stage-1 map = 100%,  reduce = 18%, Cumulative CPU 28.45 sec
2018-09-17 19:06:24,672 Stage-1 map = 100%,  reduce = 19%, Cumulative CPU 41.58 sec
2018-09-17 19:06:36,482 Stage-1 map = 100%,  reduce = 20%, Cumulative CPU 53.84 sec
2018-09-17 19:07:16,082 Stage-1 map = 100%,  reduce = 100%, Cumulative CPU 14.69 sec
MapReduce Total cumulative CPU time: 14 seconds 690 msec
Ended Job = job_1536748924580_2024 with errors
Error during job, obtaining debugging information...
Examining task ID: task_1536748924580_2024_m_000000 (and more) from job job_1536748924580_2024
Examining task ID: task_1536748924580_2024_r_000003 (and more) from job job_1536748924580_2024
Examining task ID: task_1536748924580_2024_r_000001 (and more) from job job_1536748924580_2024

Task with the most failures(10):

---

---

Diagnostic Messages for this Task:
Error: java.lang.RuntimeException: Hive Runtime Error while closing operators: null
        at org.apache.hadoop.hive.ql.exec.mr.ExecReducer.close(ExecReducer.java:286)
        at org.apache.hadoop.mapred.ReduceTask.runOldReducer(ReduceTask.java:453)
        at org.apache.hadoop.mapred.ReduceTask.run(ReduceTask.java:392)
        at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:170)
        at java.security.AccessController.doPrivileged(Native Method)
        at javax.security.auth.Subject.doAs(Subject.java:422)
        at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1866)
        at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:164)
Caused by: java.lang.NullPointerException
        at hivemall.topicmodel.ProbabilisticTopicModelBaseUDTF.finalizeTraining(ProbabilisticTopicModelBaseUDTF.java:277)
        at hivemall.topicmodel.ProbabilisticTopicModelBaseUDTF.close(ProbabilisticTopicModelBaseUDTF.java:270)
        at org.apache.hadoop.hive.ql.exec.UDTFOperator.closeOp(UDTFOperator.java:145)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:620)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:634)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:634)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:634)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:634)
        at org.apache.hadoop.hive.ql.exec.mr.ExecReducer.close(ExecReducer.java:278)
        ... 7 more

 

 

 

JAR :hivemall-all-0.5.0-incubating.jar

Issue Links

• duplicates: HIVEMALL-218 Fixed train_lda NPE where input row is null
• is related to: HIVEMALL-91 Implement Online LDA
• links to: GitHub Pull Request #165
• links to: GitHub Pull Request #166