Textbook of Linux

head prints the first lines (default 10) of one or more files or of stdin. It is the quick way to peek at a file without opening it, and the standard way to limit a pipeline's output to a manageable size.

head file.txt                      # first 10 lines
head -n 20 file.txt                # first 20 lines
head -n 1 file.txt                 # just the first line
head -c 100 file                   # first 100 bytes
head -n -5 file                    # all but the last 5 lines (GNU)
tail -n +2 file                     # skip first line (complement)
sort -rn big.log | head             # top 10

head is often paired with tail for extracting a specific range: head -n 20 file | tail -n 10 prints lines 11–20. For more surgical selection, sed -n '11,20p' file does the same thing in one step.

Piping long-running commands through head can also act as a safety valve—find / | head -n 100—terminating the upstream command once it has produced enough output (via SIGPIPE).