gerritjvv / elephant-bird Goto Github PK

merge the proto input, output pig formats.

The loader throws an exception if its passes a 0 length file.

The exception received is:

org.apache.pig.backend.executionengine.ExecException: ERROR 6018: Error while reading input
at com.twitter.elephantbird.pig.proto.LzoProtobuffB64LinePigStore.getNext(Unknown Source)
at org.apache.pig.backend.hadoop.executionengine.mapReduceLayer.PigRecordReader.nextKeyValue(PigRecordReader.java:187)
at org.apache.hadoop.mapred.MapTask$NewTrackingRecordReader.nextKeyValue(MapTask.java:423)
at org.apache.hadoop.mapreduce.MapContext.nextKeyValue(MapContext.java:67)
at org.apache.hadoop.mapreduce.Mapper.run(Mapper.java:143)
at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:621)
at org.apache.hadoop.mapred.MapTask.run(MapTask.java:305)
at org.apache.hadoop.mapred.Child.main(Child.java:170)
Caused by: java.lang.ArrayIndexOutOfBoundsException: -17
at org.apache.commons.codec.binary.Base64.isBase64(Base64.java:137)
at org.apache.commons.codec.binary.Base64.discardNonBase64(Base64.java:478)
at org.apache.commons.codec.binary.Base64.decodeBase64(Base64.java:374)
at org.apache.commons.codec.binary.Base64.decode(Base64.java:220)
... 8 more

The fix is to check for the Text.length before reading.

LzoProtobuffB64LinePigStorage.getNext line 169 is throwing an exception when it can transform the gpb line.
This causes the whole job to fail. Its more important that the job does not fail. So I've added in a try catch that writes a BADGPB counter and still allows the job to continue.

Hi Gerrit,

I have asked at kevinweil/elephant-bird whether they are interested in merging your LzoPigStorage class into their official repository:

Second, gerritjvv forked Dmitriy's EB repo some time ago (gerritjvv/elephant-bird).
He added the class LzoPigStorage.java. This class appears to be a very simple,
straight-forward LZO-enabled wrapper for standard PigStorage -- less than 10 lines
of effective code.

Dmitriy (https://github.com/dvryaboy) replied:

We can certainly pull in Gerrit's LzoPigStorage, he just needs to send a pull request :).

Would this be ok for you?

Best,
Michael

Some fields go missing when using a complex message structure. This is due to the column pruning not working correctly.