{scrollbar}

How to use Pentaho MapReduce to transform and summarize detailed data into an aggregate dataset.

It is a common use case when preparing data for extraction to an RDBMS-based data warehouse or mart. You will use parsed weblog data as the details and build an aggregate file containing a count of page views by IP address and month.

The steps in this guide include

  1. Loading the sample data file into HDFS
  2. Developing a PDI transformation which will serve as a Mapper
  3. Developing a PDI transformation which will serve as a Reducer
  4. Developing a PDI job which will invoke a Pentaho MapReduce step that runs MapReduce using the developed mapper and reducer transformation.
  5. Executing and reviewing output

Prerequisites

In order follow along with this how-to guide you will need the following:

Sample Files

The sample data file needed for this guide is:

File Name

Content

weblogs_parse.txt.zip

Tab-delimited, parsed weblog data


NOTE: If you have completed the Using Pentaho MapReduce to Parse Weblog Data guide, then the necessary files will already be in the proper location.
This file should be placed into HDFS at /user/pdi/weblogs/parse using the following commands:

hadoop fs –mkdir /user/pdi/weblogs
hadoop fs –mkdir /user/pdi/weblogs/parse
hadoop fs –put weblogs_parse.txt /user/pdi/weblogs/parse/

Step-By-Step Instructions

Setup

Start Hadoop if it is not already running.

Create a Job to Execute a MapReduce Process

In this task you will create a job that will execute a MapReduce process that runs the newly created mapper and reducer transformations.

  1. Start PDI on your desktop. Once it is running choose 'File' -> 'New' -> 'Job' from the menu system or click on the 'New file' icon on the toolbar and choose the 'Job' option.

    You can download the Kettle Job aggregate_mr.kjb already completed



  2. Add a Start Job Entry: You need to tell PDI where to start the job, so expand the 'General' section of the Design palette and drag a 'Start' node onto the job canvas. Your canvas should look like:


  3. Add a Pentaho Map Reduce Job Entry: You are creating the job to execute a Pentaho MapReduce transformation, so expand the 'Big Data' section of the Design palette and drag a 'Pentaho MapReduce' node onto the job canvas. Your canvas should look like:


  4. Connect the Start and MapReduce Job Entries: Hover the mouse over the 'Start' node and a tooltip will appear. Click on the output connector (the green arrow pointing to the right) and drag a connector arrow to the 'Pentaho MapReduce' node.

    Your canvas should look like this:


  5. Edit the MapReduce Job Entry: Double-click on the 'Pentaho MapReduce' node to edit its properties. Enter this information:
    1. Hadoop Job Name: Enter 'Aggregate Map Reduce'
    2. Mapper Transformation: Enter <PATH>/aggregate_mapper.ktr
      <PATH> is the folder path you saved the mapper in.
    3. Mapper Input Step Name: Enter 'Map/Reduce Input'
    4. Mapper Output Step Name: Enter 'Map/Reduce Output'
      When you are done the window should look like:

  6. Configure the Reducer: Switch to the 'Reducer' tab and enter the following:
    1. Reducer Transformation: Enter <PATH>/aggregate_reducer.ktr
    2. Reducer Input Step Name: Enter 'Map/Reduce Input'
    3. Reducer Output Step Name: Enter 'Map/Reduce Output'
      When you are done the window should look like:


  7. Configure the MapReduce Job: Switch to the 'Job Setup' tab. Enter this information:
    1. Input Path: Enter '/user/pdi/weblogs/parse'
    2. Output Path: Enter '/user/pdi/weblogs/aggregate_mr'
    3. Input Format: Enter 'org.apache.hadoop.mapred.TextInputFormat'
    4. Output Format: Enter 'org.apache.hadoop.mapred.TextOutputFormat' 
    5. Check 'Clean output path before execution'
      When you are done your window should look like:


  8. Configure the Cluster Properties: Switch to the 'Cluster' tab. Enter this information:
    1. Hadoop distribution: Select your Hadoop distribution
    2. Working Directory: Enter '/tmp'
    3. HDFS Hostname, HDFS Port, Job Tracker Hostname, Job Tracker Port: Your connection information.
    4. Number of Mapper Tasks: Enter '3'. You can play around with this to get the best performance based on the size of your data and the number of nodes in your cluster.
    5. Number of Reducer Tasks: Enter '1'. You can play around with this to get the best performance based on the size of your data and the number of nodes in your cluster.
    6. Check 'Enable Blocking'
    7. Logging Interval: Enter '10'. The number of seconds between pinging Hadoop for completion status messages
      When you are done your window should look like:

      Click 'OK' to close the window.

  9. Save the Job: Choose 'File' -> 'Save as...' from the menu system. Save the transformation as 'aggregate_mr.kjb' into a folder of your choice.

  10. Run the Job: Choose 'Action' -> 'Run' from the menu system or click on the green run button on the job toolbar. A 'Execute a job' window will open. Click on the 'Launch' button. An 'Execution Results' panel will open at the bottom of the PDI window and it will show you the progress of the job as it runs. After a few seconds the job should finish successfully:

    If any errors occurred the job step that failed will be highlighted in red and you can use the 'Logging' tab to view error messages.

Check Hadoop for Aggregated Web Log

  1. Run the following command to view the aggregated results:
    hadoop fs -cat /user/pdi/weblogs/aggregate_mr/part-00000 | head
    
    

    This should return the first few rows of the aggregated file.

Summary

During this guide you learned how to create and execute a Pentaho MapReduce job on a Hadoop cluster. You consumed detailed weblog data and generated an aggregate datafile which is suitable for load into an RDBMS-based data warehouse or mart.

{scrollbar}