Overview
head prints the beginning of each file you name, ten lines by default. Use it to peek at a large file or to keep the first results of a pipeline.
Background
The head command appeared in PWB Unix, the Programmer's Workbench, according to the FreeBSD and macOS manuals. Today the name covers GNU head, the head that macOS ships, and the head applet in BusyBox. They agree on -n and -c with a positive count, and differ on negative counts and header options.
Examples
Print the first ten lines (all of a short file):
head notes.txt
Print the first two lines:
head -n 2 notes.txt
Print the first line of several files; each is introduced by a header:
head -n 1 app.log logs/server.log
Print the first five bytes:
head -c 5 notes.txt
Keep only the first two lines of a pipeline's output:
printf '1\n2\n3\n4\n' | head -n 2
Options
-n
Print this many lines instead of ten. The count is the next argument, as in -n 2.
-c
Print this many bytes instead of lines. A multibyte character can be cut in half.
Common mistakes
A count like -n 2 is a number of lines, not a line number. When several files are given, head prints a ==> name <== header for each; use -q where supported to hide it.
Portability
Only -n and -c with a positive count are in the Common examples. The Common examples were executed on GNU coreutils 9.11, macOS system tools, and BusyBox 1.37.0 (see the development guide for the exact check).
About macOS head
This is the head in /usr/bin on macOS. It does not accept a negative count, so use tail or sed to drop trailing lines.
Apple notes
macOS head is the BSD implementation. It does not accept a negative count such as -n -2; use tail or sed to drop trailing lines.
References
Original explanations, with upstream documentation for further reading. FreeBSD references are comparison material, not a claim of Apple version equivalence.