The "data wrangling" chapter goes off the rails around the time it starts massaging the regex to handle the edge case "What if the user's username is 'Disconnec

The "data wrangling" chapter goes off the rails around the time it starts massaging the regex to handle the edge case "What if the user's username is 'Disconnected from'?" In my experience if you are using sed or awk on the logs you should give in and write some Python, or start using an actual log tool like Splunk. You'll never remember the awk syntax anyway. On the other hand, the idea of piping your log data to a statistics package or gnuplot is fascinating. It could have benefited from showing some output from those programs to illustrate the kind of things you can get from it.