NSE disseminates files containing `snapshots' of the limit order book, at various points of time every day. These files are just flat ascii files, with millions of lines which look like --
200108310026586|SRGINFOTEC|EQ|5000|0.80|09:58:41|B|ynnn|nnn|nnn|RL|0|0|. 200108310029204|SRGINFOTEC|EQ|1000|0.80|09:59:07|B|ynnn|nnn|nnn|RL|0|0|. 200108310036557|SRGINFOTEC|BE|500|0.80|10:00:19|B|ynnn|nnn|nnn|RL|0|0|. 200108310044134|SRGINFOTEC|BE|1000|0.80|10:01:40|B|ynnn|nnn|nnn|RL|0|0|. 200108310044630|SCINTSOFT|BE|5000|0.80|10:01:46|S|ynnn|nnn|nnn|RL|0|0|. 200108310046098|NIPPONDENR|EQ|6364|0.80|10:02:02|B|ynnn|nnn|nnn|RL|0|0|.
Useful Fact: All the snapshots for August 2001 compress to 64,076 megabytes using bzip2. They compress to 89,535 megabytes using gzip. In both cases, no commandline args were used. Thus the space taken by bzip2 is 71.565% of that taken by gzip, for this problem.
Ajay Shah, 2001