Shows how different encoding choices can make big speed differences.
Run with command line options --generate 1000000 test.csv to generate a million data lines in
test.csv.
Run with command line options --parser test.csv to time how long it takes to parse and encode
those million data points
Run with command line options --fast test.csv to time how long it takes to parse and encode those
million data points using byte-level parsing and direct value encoding.
This doesn't demonstrate text encoding which is subject to somewhat different tricks. The basic
idea of caching hash locations and byte level parsing still very much applies to text, however.