Java MapReduce with Hadoop

MapReduce is a “corset” and forces the developer into narrow boundaries. Therefore, it makes sense to read “MapReduce Design Patterns” to quickly learn the common tricks and techniques. It is similar to learning other paradigms, such as Divide-and-Conquer or recursion. You grasp the general principle quickly, but you have to learn when and how to best use it.

In this book, typical use cases for Hadoop are implemented in Java MapReduce. The focus is not on the actual design patterns, as the title of the book suggests, but on the concrete implementations. Anyone who is only interested in the implementation of these patterns will be well served by this book. However, you should already have some initial experience with Hadoop and know Java well.

However, anyone looking for real “design patterns” for arbitrary MapReduce frameworks will not be 100% happy with this book. The “other systems” mentioned in the subtitle definitely fall short. Also, those who use Hadoop with Pig, Hive, or Apache Crunch might be able to skip reading the book or just skim through it. You can then save yourself the trouble of going through the long Java listings.

However, for those who really want to know what happens behind Pig, whether for troubleshooting or performance issues, you can certainly pick up a tip or two here.

• Donald Miner, Adam Shook
• MapReduce Design Patterns: Building Effective Algorithms and Analytics for Hadoop and Other Systems
• O’Reilly
• 2012

See also the review on Amazon.