and also load from s3
data = load 's3n://...'
The "n" stands for native. I believe S3 also supports block device storage (s3://) which allows bigger files to be stored. I don't know how (if at all) the two types interact.
On Apr 7, 2013, at 1:11 PM, Panshul Whisper <[EMAIL PROTECTED]> wrote:
> I am trying to run a pig script which is suppoesed to read input from s3
> and write back to s3. The cluster
> scenario is as follows:
> * Cluster is installed on EC2 using Cloudera Manager 4.5 Automatic
> * Installed version: CDH4
> * Script location on - one of the nodes of cluster
> * running as : $ pig countGroups_daily.pig
> *The Pig Script*:
> set fs.s3.awsAccessKeyId xxxxxxxxxxxxxxxxxx
> set fs.s3.awsSecretAccessKey xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> --load the sample input file
> data = load 's3://steamdata/nysedata/NYSE_daily.txt' as
> (exchange:chararray, symbol:chararray, date:chararray, open:float,
> high:float, low:float, close:float, volume:int, adj_close:float);
> --group data by symbols
> symbolgrp = group data by symbol;
> --count data in every group
> symcount = foreach symbolgrp generate group,COUNT(data);
> --order the counted list by count
> symcountordered = order symcount by $1;
> store symcountordered into 's3://steamdata/nyseoutput/daily';
> Message: org.apache.pig.backend.executionengine.ExecException: ERROR 2118:
> Input path does not exist: s3://steamdata/nysedata/NYSE_daily.txt
> Failed to read data from "s3://steamdata/nysedata/NYSE_daily.txt"
> Please help me, what am I doing wrong. I can assure you that the input
> path/file exists on s3 and the AWS key and secret key entered are correct.
> Thanking You,
> Ouch Whisper