Efficient column based data encoding for large-scale data storage

摘要: The subject disclosure relates to column based data encoding where raw be compressed is organized by columns, and then, as first second layers of reduction the size, dictionary and/or value are applied create integer sequences that correspond columns. Next, a hybrid greedy run length bit packing compression algorithm further compacts according an analysis savings. Synergy techniques in concert with column-based organization, coupled gains scanning querying efficiency owing representation compact data, results substantially improved at fraction cost conventional systems.