Fig. 1From: FASTA/Q data compressors for MapReduce-Hadoop genomics: space and time savings made easyThe layout of a block-oriented compressed data file when uploaded to HDFS. In the figure, a the original file includes an header, a footer and 8 compressed data blocks. b When uploaded to HDFS, it is partitioned into 4 HDFS data blocks. c As a result of the partitioning, the compressed data block labeled as CB5 is divided into two parts and assigned to two different HDFS data blocks. Using the Compressed Block Split strategy, each compressed data block is modeled as a distinct split. d Using the Enhanced Split strategy, several compressed data blocks are grouped into fewer input splitsBack to article page