Part III · Shell and commands

28.2 grep and regex

alsogrepregexregular expressionregexpegreprgripgrepsearch text

TL;DR

grep prints the lines that match a pattern, in files or in what you pipe into it.
-r searches a whole folder, -i ignores case, -n shows line numbers, -v keeps the lines that don't match.
-E for the usual regex syntax (+, ?, |, ()), -F to search a plain string.

grep -rn 'TODO' src/
journalctl -b | grep -i 'error'
grep -E 'nginx|php-fpm' services.txt

Pitfalls

Put the pattern in single quotes, or the shell may change *, $ or ! before grep sees it.
Without -E, | and + match themselves: grep 'a|b' looks for the three characters a|b.
. matches any character: grep '10.0.0.1' also matches 10a0b0c1. Use -F for a literal string.
ps aux | grep nginx also lists the grep itself: pgrep -a nginx doesn't.
grep -r reads binary files too: add -I to skip them. It doesn't follow symbolic links; -R does.

Deep dive 2 questions

What does a regex look like?

^ is the start of the line and $ its end. . is any character, [abc] one of these, [^0-9] anything but a digit. * repeats what comes before zero or more times. With -E: + once or more, ? optional, {2,4} between 2 and 4 times, a|b either one, and ( ) groups.

grep -E '^[0-9]{3}-[0-9]{4}$' numbers.txt

grep, egrep or ripgrep?

egrep and fgrep are old names for grep -E and grep -F; since version 3.8, GNU grep warns when you use them. rg (ripgrep, package ripgrep) is a faster alternative: it searches recursively by default and skips hidden files, binary files and anything in .gitignore.

Commands: grep rg
See also: 28.1 find
28.4 sed