Reputation: 1063
My sequence files are stored directly in hdfs e.g.:
grunt> ls
grunt> ls /blabla
hdfs://namenode1:54310/blabla/0411f03a-db7f-48d0-9542-5203304e3e81.seq<r 3> 185284523
hdfs://namenode1:54310/blabla/05be8fc0-e967-42e1-b76a-0d7108a69d17.seq<r 3> 201489688
hdfs://namenode1:54310/blabla/06222427-519c-49c0-bbbf-49a9f43bbd13.seq<r 3> 196858576
hdfs://namenode1:54310/blabla/066da26a-48da-45b1-83f5-60d16475e40d.seq<r 3> 194832641
hdfs://namenode1:54310/blabla/07cbfc83-42a2-47bf-b364-d39da3a2d071.seq<r 3> 194806047
hdfs://namenode1:54310/blabla/10dea7b8-9ed3-4e66-b4bd-a3c07d8bf39e.seq<r 3> 166224702
How can I create a Pig script which is reading every file from the directory "blabla" and performing an action?
I've tried multiple ways for loading the input but none of those worked:
%default INPUT '/blabla/f8fbbe9a-aae3-413f-b3b9-37cdef71da8f.seq'
%default INPUT 'hdfs://namenode1:54310/blabla/f8fbbe9a-aae3-413f-b3b9-37cdef71da8f.seq'
%default INPUT 'f8fbbe9a-aae3-413f-b3b9-37cdef71da8f.seq'
I always get the error:
Input(s):
Failed to read data from "hdfs://namenode1:54310/........."
Upvotes: 1
Views: 1817
Reputation: 449
You can try reading the Sequence Files in these ways :
Pig SequenceFileLoader :
A = LOAD 'hdfs://namenode1:54310/blabla/*' using org.apache.pig.piggybank.storage.SequenceFileLoader();
(Or) Using Elephant Bird :
REGISTER 'elephant-bird-pig-3.0.5.jar';
REGISTER 'elephant-bird-core-4.1.jar';
REGISTER 'elephant-bird-hadoop-compat-4.1.jar';
A = LOAD 'hdfs://namenode1:54310/blabla/*' using com.twitter.elephantbird.pig.load.SequenceFileLoader();
Upvotes: 2
Reputation: 403
Did you try this way :
%default INPUT 'hdfs://namenode1:54310/blabla/*'
?
It should work if your .seq files are readables. It looks like they are not, because your attempt to do it should have load one file. Could-you give the complete log line?
Maybe you would have to use pig SequenceFileLoader.
Upvotes: 0