Today during a bunch of bioinformatics analysis task, I had to obtain a bunch of lines randomly from the file. Guess what? My instinct was correct. There exists a nice tool exactly for this task provided in GNU tools package
shuf.
Using shuf, to get N random lines:
shuf -n N input > output
Another way is to use sort tool to sort the file randomly and then pick the
first N lines. For example:
sort -R input | head -n N > output
Awesome!
A more fruitful discussion with respect to efficiency of different tools can be found on StackOverflow.