[
https://issues.apache.org/jira/browse/PIG-1917?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13017219#comment-13017219
]
Thejas M Nair commented on PIG-1917:
------------------------------------
To avoid breaking backward compatibility while supporting this use case, one
option would be to have a way to switch to the mode that supports this use case.
Maybe by adding a " using 'unquote-mode' " to the command ?
eg .
{code}
B = MAPREDUCE 'hadoop-examples.jar' Store A into 'inputDir' Load 'outputDir' as
(word:chararray, count:int)
`wordcount -Dmapred.job.map.memory.mb=3072 -Dmapred.child.java.opts="-Xmx1536m
-Xss128k" inputDir outputDir` using 'unquote-mode';
{code}
or maybe by setting a system property on commandline -
-Dpig.unquote.cmdstring='true'
I think adding a system property is better, as it does not pollute the syntax.
In 0.9, maybe we can make the unquote mode the default mode.
> NativeMapReduce does not Allow Configuration Parameters containing Spaces
> --------------------------------------------------------------------------
>
> Key: PIG-1917
> URL: https://issues.apache.org/jira/browse/PIG-1917
> Project: Pig
> Issue Type: Bug
> Reporter: Mitesh Singh Jat
> Assignee: Thejas M Nair
> Fix For: 0.8.0
>
>
> In Pig, NativeMapReduce does not allow Configuration parameters which contain
> spaces.
> Considering a simple pig script with _wordcount_ as Native M-R job.
> {code:title=nativeMR.pig|borderStyle=solid}
> A = load 'input/WordCountInput.txt';
> B = MAPREDUCE 'hadoop-examples.jar' Store A into 'inputDir' Load 'outputDir'
> as (word:chararray, count:int) `wordcount -Dmapred.job.map.memory.mb=3072
> -Dmapred.child.java.opts="-Xmx1536m -Xss128k" inputDir outputDir`;
> store B into 'output/WordCount' using PigStorage;
> {code}
> The above pig script fails while launching native MAPREDUCE job. Because
> space after _-Dmapred.child.java.opts="-Xmx1536m_ makes
> _-Xss128k"_ as the first argument to _wordcount_
> Even with other example M-R program, the parameter _mapred.child.java.opts_
> was assigned _"-Xmx1536m_ in Jobconf.
> Physical plan, Logical plan, and M-R plan of pig show double quotes.
> For example, here is the corresponding M-R Plan for the native M-R
> "MAPREDUCE" statement.
> {code}
> MapReduce node scope-13
> hadoop jar hadoop-examples.jar wordcount -Dmapred.job.map.memory.mb=3072
> -Dmapred.child.java.opts="-Xmx1536m -Xss128k" inputDir outputDir
> --------
> {code}
> On executing above M-R job after copying 'input/WordCountInput.txt' into
> 'inputDir/', as Hadoop Job, this ran successfully.
> {code}
> $ *hadoop jar hadoop-examples.jar wordcount -Dmapred.job.map.memory.mb=3072
> -Dmapred.child.java.opts="-Xmx1536m -Xss128k" inputDir outputDir*
> 11/03/18 14:23:29 INFO input.FileInputFormat: Total input paths to process : 1
> 11/03/18 14:23:29 INFO mapred.JobClient: Running job: job_201103181353_0007
> 11/03/18 14:23:30 INFO mapred.JobClient: map 0% reduce 0%
> 11/03/18 14:23:37 INFO mapred.JobClient: map 100% reduce 0%
> 11/03/18 14:23:49 INFO mapred.JobClient: map 100% reduce 100%
> 11/03/18 14:23:51 INFO mapred.JobClient: Job complete: job_201103181353_0007
> ...
> ...
> {code}
> Whereas, when _-Dmapred.child.java.opts=-Xmx1536m_ was used, the
> *nativeMR.pig* executed successfully.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
