Track · 1:26 · Liner note

Compression Ratio

Compression ratio is the original size of data divided by its compressed size, and in columnar databases similar, well-sorted values compress especially well, cutting storage and I/O.

A storage graph that goes the right way is a real pleasure. Compression ratio measures it: how many bytes you started with against how many you keep. Similar values stored together compress well, which is why column order and data types are design decisions, not housekeeping.

This one fits the afternoon you re-sort a table, rerun the size query and watch the footprint fall. Less disk, less I/O, faster scans. Elegance, it turns out, is often just redundancy removed.