猪udf中python代码的正确输入/输出?

时间:2013-03-29 23:01:39

标签: python hadoop apache-pig

我有这个简短的python脚本:

import langid
import sys

for pig_tuple in sys.stdin:
    cols = pig_tuple.split()

    if len(cols) < 2:
        sys.exit(0)

    try:
        id = int(cols[0])
        text = " ".join(cols[1:])
    except:
        sys.exit(0)

    (lang,prob) = langid.classify(text)
    print "%s\t%s" %(id,lang)

sys.exit(0)

我想在猪脚本中运行它。我试过了:

define langid_cmd `python2.6 /data/test/compiled_python/langid_command_line.py` ship('/data/test/compiled_python/langid_command_line.py');

text = LOAD '$PIG_INPUT' USING PigStorage() as (text:chararray);

pythonDetect1 = STREAM text through langid_cmd AS (pid:chararray,planguage:chararray);

但我明白了:

2013-03-29 15:53:22,290 [main] INFO  org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.MapReduceLauncher - Failed!
2013-03-29 15:53:22,303 [main] ERROR org.apache.pig.tools.grunt.GruntParser - ERROR 2999: Unexpected internal error. java.lang.String cannot be cast to org.apache.pig.data.Tuple
Details at logfile: /home/isl/ryan/src/main/pigScripts/pig_1364597410350.log
2013-03-29 15:53:22,306 [main] ERROR org.apache.pig.tools.grunt.GruntParser - ERROR 2999: Unexpected internal error. java.lang.String cannot be cast to org.apache.pig.data.Tuple
Details at logfile: /home/isl/ryan/src/main/pigScripts//src/main/pigScripts/pig_1364597410350.log
2013-03-29 15:53:22,308 [main] ERROR org.apache.pig.tools.grunt.GruntParser - ERROR 2999: Unexpected internal error. java.lang.String cannot be cast to org.apache.pig.data.Tuple
Details at logfile: /home/isl/ryan/src/main/pigScripts//src/main/pigScripts/pig_1364597410350.log
2013-03-29 15:53:22,311 [main] ERROR org.apache.pig.tools.grunt.GruntParser - ERROR 2999: Unexpected internal error. java.lang.String cannot be cast to org.apache.pig.data.Tuple
Details at logfile: /home/isl/ryan/src/main/pigScripts/src/main/pigScripts/pig_1364597410350.log
2013-03-29 15:53:22,313 [main] ERROR org.apache.pig.tools.grunt.GruntParser - ERROR 2999: Unexpected internal error. java.lang.String cannot be cast to org.apache.pig.data.Tuple
Details at logfile: /home/isl/ryan/src/main/pigScripts/src/main/pigScripts/pig_1364597410350.log

目录/ data / test / compiled_python是chmod到777,当我从shell运行时:

-bash-3.2$ echo 14353 I can haz pigscriptz? | python /data/test/compiled_python/langid_command_line.py 
14353   eu

...

1 个答案:

答案 0 :(得分:0)

AS (pid:chararray,planguage:chararray)告诉pig期望输出是一个字符串元组,但是你返回制表符分隔的字符串。您应该将结果打印出来

print "(%s,%s)" %(id,lang)

or use the python UDF integration