Previously, UC Berkeley’s Shark had been giving me build fits. Finally tamed the savage beast. Had to resort to java -verbose:class to figure out where various classes and jars were coming from. Then a little surgery to put the appropriate Hadoop jars in the proper place and voilà, I can run the Shark examples against an hdfs repository.
With that out of the way, I’m now getting sucked into Apache Hive development. Yum, data crunching for fun and profit. I didn’t read too closely but I like what I perused of Programming Hive by Dean Wampler. Feels comprehensive and up to date.