我的建议：

1. 在hdfs中创建一个文件夹： hadoop fs -mkdir / pigdata $>


2. 加载文件创建的hdfs文件夹： hadoop fs -put /opt/pig/tutorial/data/excite-small.log/ pigdata




（或者你可以从grunt shell作为 grunt> copyFromLocal /opt/pig/tutorial/data/excite-small.log / pigdata ）

1. 执行pig latin脚本：
   
   

     grunt>在
   
    grunt>上设置调试set job.name'first-p2-job'
   
    grunt> log = LOAD'hdfs：// hostname：54310 / pigdata / excite-small.log'AS 
   （user：chararray，time：long，query：chararray）; 
    grunt> grpd = GROUP log BY用户; 
    grunt> cntd = FOREACH grpd GENERATE组，COUNT（log）; 
    grunt> STORE cntd INTO'output'; 
    




2. 输出文件将存储在 hdfs：//主机名：54310 / pigdata / output

I have a pig script, and need to load files from local hadoop cluster. I can list the files using hadoop command: hadoop fs –ls /repo/mydata,` but when i tried to load files in pig script, it failed. the load statement is like this:

in = LOAD '/repo/mydata/2012/02' USING PigStorage() AS (event:chararray, user:chararray)

the error message is:

Message: org.apache.pig.backend.executionengine.ExecException: ERROR 2118: Input path does not exist: file:/repo/mydata/2012/02

any idea? thanks

My suggestion:

1. Create a folder in hdfs : hadoop fs -mkdir /pigdata

2. Load the file to the created hdfs folder: hadoop fs -put /opt/pig/tutorial/data/excite-small.log /pigdata

(or you can do it from grunt shell as grunt> copyFromLocal /opt/pig/tutorial/data/excite-small.log /pigdata)

1. Execute the pig latin script :

      grunt> set debug on
   
      grunt> set job.name 'first-p2-job'
   
      grunt> log = LOAD 'hdfs://hostname:54310/pigdata/excite-small.log' AS 
                 (user:chararray, time:long, query:chararray); 
      grunt> grpd = GROUP log BY user; 
      grunt> cntd = FOREACH grpd GENERATE group, COUNT(log); 
      grunt> STORE cntd INTO 'output';
   

2. The output file will be stored in hdfs://hostname:54310/pigdata/output