June 2018
Intermediate to advanced
478 pages
10h 52m
English
Compression efficiency is at the heart of BLU performance. Each columnar table has a column compression dictionary made up of small symbols representing repeated bytes of data. The dictionaries are of fixed size, so there is a limited number of slots. Populating these slots with a valid representation of the data in the table is important. Text data generally compresses better than numeric and random values; those from encryption would result in poor compression results. Compression and statistics collection are automatic.
Read now
Unlock full access