SequenceFile

A SequenceFile is a flat, binary file type that serves as a container for data to be used in Apache Hadoop distributed computing projects. SequenceFiles are used extensively with MapReduce.

Since Hadoop functions best with larger files, SequenceFiles are used to store and compress files that are smaller than the optimum size for operating efficiently with Hadoop, which can help reduce required disk space capacity and I/O  requirements.

SequenceFiles serve as a container for a sequence of files. Keys are listed for reference and values, and the contents of the file are referenced in a given key. SequenceFiles support a Writer, a Reader and a Sorter class for respective functions in relation to keys. As an example, a SequenceFile might contain a massive number of log files for a server where the key would be a timestamp and the value would be the entire log file. Normally, the small text files would be very inefficient in Hadoop. After packaging into SequenceFiles, however, they can be used effectively.

Beyond packaging files into a manageable size for Hadoop, SequenceFiles support compression of the keys, the values or both.  When both are compressed, the file keys and values are collected into blocks and separately compressed. The type of compression chosen determines the file format.