python - Apache Hive 在使用 Python UDF 时出错

Question

我在 Apache 配置单元中使用 Python 用户定义函数将字符从小写字符更改为大写字符。我收到“关闭运算符时 Hive 运行时错误”的错误。

下面是我试过的查询:

describe table1;     
OK
item    string  
count   int 
city    string  

select * from table1;
aaa 1   tokyo
aaa 2   london
bbb 3   washington
ccc 4   moscow
ddd 5   bejing

从上表中，项目和城市字段应从小写变为大写，计数应增加 10。

使用的 Python 脚本:

cat caseconvert.py
import sys
import string

for line in sys.stdin:
    line = line.strip()
    item,count,city=line.split('\t')

    ITEM1=item.upper()
    COUNT1=count+10
    CITY1=city.upper()
    print '\t'.join([ITEM1,str(COUNT1),FRUIT1])

将表1数据插入表2

create table table2(ITEM1 string, COUNT1 int, CITY1 string) row format delimited fields terminated by ',';

add FILE caseconvert.py 

insert overwrite table table2 select TRANSFORM(item,count,city) using 'python caseconvert.py' as (ITEM1,COUNT1,CITY1) from table1;

如果我执行，我会收到以下错误。我无法追踪问题。我能知道它出错了吗？

Total MapReduce jobs = 3
Launching Job 1 out of 3
Number of reduce tasks is set to 0 since there's no reduce operator
Starting Job = job_201508151858_0014, Tracking URL = http://0.0.0.0:50030/jobdetails.jsp?jobid=job_201508151858_0014
Kill Command = /usr/lib/hadoop/bin/hadoop job  -kill job_201508151858_0014
Hadoop job information for Stage-1: number of mappers: 1; number of reducers: 0
2015-08-15 22:24:06,212 Stage-1 map = 0%,  reduce = 0%
2015-08-15 22:25:01,559 Stage-1 map = 100%,  reduce = 100%
Ended Job = job_201508151858_0014 with errors
Error during job, obtaining debugging information...
Job Tracking URL: http://0.0.0.0:50030/jobdetails.jsp?jobid=job_201508151858_0014
Examining task ID: task_201508151858_0014_m_000002 (and more) from job job_201508151858_0014

Task with the most failures(4): 
-----
Task ID:
  task_201508151858_0014_m_000000

URL:
  http://localhost.localdomain:50030/taskdetails.jsp?jobid=job_201508151858_0014&tipid=task_201508151858_0014_m_000000
-----
Diagnostic Messages for this Task:
java.lang.RuntimeException: Hive Runtime Error while closing operators
    at org.apache.hadoop.hive.ql.exec.ExecMapper.close(ExecMapper.java:224)
    at org.apache.hadoop.mapred.MapRunner.run(MapRunner.java:57)
    at org.apache.hadoop.mapred.MapTask.runOldMapper(MapTask.java:417)
    at org.apache.hadoop.mapred.MapTask.run(MapTask.java:332)
    at org.apache.hadoop.mapred.Child$4.run(Child.java:268)
    at java.security.AccessController.doPrivileged(Native Method)
    at javax.security.auth.Subject.doAs(Subject.java:396)
    at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1438)
    at org.apache.hadoop.mapred.Child.main(Child.java:262)
Caused by: org.apache.hadoop.hive.ql.metadata.HiveException: [Error 20003]: An error occurred when trying to close the Operator running your custom script.
    at org.apache.hadoop.hive.ql.exec.ScriptOperator.close(ScriptOperator.java:488)
    at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:570)
    at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:5

FAILED: Execution Error, return code 20003 from org.apache.hadoop.hive.ql.exec.MapRedTask. An error occurred when trying to close the Operator running your custom script.
MapReduce Jobs Launched: 
Job 0: Map: 1   HDFS Read: 0 HDFS Write: 0 FAIL
Total MapReduce CPU Time Spent: 0 msec

Answer 1

在 Python 脚本的最后一行，将输出打印到 STDOUT 的地方，您调用了 FRUIT1 而没有定义它。这应该是 CITY1。您还导入了字符串但未使用它。我会以不同的方式编写脚本:

import sys
import string

while True:
    line = sys.stdin.readline()
    if not line:
        break

    line = string.strip(line, '\n ')
    item,count,city=string.split(line, '\t')

    ITEM1=item.upper()
    COUNT1=count+10
    CITY1=city.upper()
    print '\t'.join([ITEM1,str(COUNT1),CITY1])

然后，我将使用 CREATE TABLE AS SELECT 查询(假设 TABLE1 和您的 python 脚本都在 HDFS 中):

create table TABLE2
as select transform(item, count, city)
using 'hdfs:///user/username/caseconvert.py' 
as (item1 string, count1 string, city1 string)
FROM TABLE1;

这对我有用。但是，有一种更简单的方法可以使用 Hive 内置函数进行所需的转换:

upper(string A) >>> 返回将 A 的所有字符转换为大写后的字符串。例如，upper('fOoBaR') 结果为 'FOOBAR'。

当然，对于城市，您可以做到:(city + 10) AS city1。

因此，可以按如下方式创建 TABLE2:

CREATE TABLE2
AS SELECT
UPPER(ITEM) AS ITEM1,
COUNT + 10 AS COUNT1,
UPPER CITY AS CITY1
FROM TABLE1;

比编写自定义 UDF 麻烦得多。