Flexible In-Situ Indexing for Hadoop via Elephant TwinThis document discusses flexible indexing in Hadoop. It describes how Twitter uses Elephant-Twin, an open source library they developed, to create indexes at the block level or record level in Hadoop. Elephant-Twin allows minimal changes to jobs/scripts, indexes data without copying it, supports post-factum indexing, and indexes can be used to efficiently retrieve relevant data through an IndexedI
