our input is a line of text which may be parsed to e.g. A or B object. We want all A objects written to "A.avro" files, while all B objects written to "B.avro".
I looked into AvroMultipleOutputs class: http://avro.apache.org/docs/1.7.4/api/java/org/apache/avro/mapreduce/AvroMultipleOutputs.html There is an example, however, it's not quite clear. For job submission, it uses AvroMultipleOutputs.addNamedOutput to add schemas for A and B. In my program looks like: AvroMultipleOutputs.addNamedOutput(job, "A", AvroKeyOutputFormat.class, aSchema, null); AvroMultipleOutputs.addNamedOutput(job, "B", AvroKeyOutputFormat.class, bSchema, null); I believe this is for Reducer output files.
*My question is* what the Mapper output should be, in specific what "job.setMapOutputValueClass" should be, since the Mapper output could be A or B object, with schema aSchema or bSchema.
In my progam, I simply set it to GenericData, but get error as below:
14/03/06 15:55:34 INFO mapreduce.Job: Task Id : attempt_1393817780522_0012_m_000010_2, Status : FAILED Error: java.lang.NullPointerException at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.init(MapTask.java:989) at org.apache.hadoop.mapred.MapTask.createSortingCollector(MapTask.java:390) at org.apache.hadoop.mapred.MapTask.access$100(MapTask.java:79) at org.apache.hadoop.mapred.MapTask$NewOutputCollector.<init>(MapTask.java:674) at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:746) at org.apache.hadoop.mapred.MapTask.run(MapTask.java:339) at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:165) at java.security.AccessController.doPrivileged(Native Method) at javax.security.auth.Subject.doAs(Subject.java:415) at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1491) at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:160)
any idea on how to build a common map output data type? The only way I can think of is "toString()", which would be very inefficient, since A and B are big objects and may change with time, which is also the reason we want to use Avro serialization. 2014-03-07 9:55 GMT+08:00 Harsh J <[EMAIL PROTECTED]>:
You may consider "SpecificRecord" or "GenericRecord" of Avor. Yong
Date: Fri, 7 Mar 2014 10:29:49 +0800 Subject: Re: MapReduce: How to output multiplt Avro files? From: [EMAIL PROTECTED] To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
thanks, Harsh. any idea on how to build a common map output data type? The only way I can think of is "toString()", which would be very inefficient, since A and B are big objects and may change with time, which is also the reason we want to use Avro serialization. 2014-03-07 9:55 GMT+08:00 Harsh J <[EMAIL PROTECTED]>:
If you have a reducer involved, you'll likely need a common map output
data type that both A and B can fit into.
On Thu, Mar 6, 2014 at 12:09 AM, Fengyun RAO <[EMAIL PROTECTED]> wrote:
All projects made searchable here are trademarks of the Apache Software Foundation.
Service operated by Sematext