`sed`, `awk`, `grep`, and the like - cheatsheet

sed and awk both read from stdin (or can be given a filename as an argument) and print to stdout.

`sed`

Command	Function
`sed s/foo/bar/g file.txt`	Replace `foo` (regexp) with `bar` in `file.txt`.
`sed 5 d`	Delete line 5.
`sed /foo/d`	Delete lines matching `foo`.
`sed -n 5,30 p`	Print lines 5 to 30.
`sed -n s/foo/bar/p`	Print only changed lines.
`sed i 17 foo`	Insert "foo" at line 17.
`sed /foo/a bar`	Append "bar" after lines containing "foo".
`sed s#foo#bar#g`	You don't have you use '/'s for regexes.
`seg G`	Double space input.
`sed -e "/12:00/,/12:10/!d" full.log > wanted.log`	Top 'n' tail a log file by time stamp.

Option	Effect
-n	Supress normal output - use `p` to print
-i	Replace file in place. BSD sed requires a suffix.

Basic structure:

BEGIN {ACTION}
CONDITION {ACTION}
CONDITION {ACTION}
END {ACTION}

Command	Function
`awk -F, '{print $5}'`	Split on commas, print 5th token. `$NF` would give you the last.
`awk 'BEGIN {s = 0} {s += $3} END {print s}'`	Sum 3rd column.
`awk '/foo/ {print $0}'`	Print lines matching `foo`. `$0` is the whole line.
`awk '/foo/'`	As above - `print` is the default action.
`awk 'length($0)'`	Pring long lines.
`awk '/a/ && /b/ && !/c/ && /d/'`	Chained expressions - equivelent to `grep a

Command	Function
`grep foo`	Include only lines condaining `foo`.
`grep -v foo`	Exclide lines containing `foo`.
`grep -o '{.*}'`	Include only part of lines matching pattern. (This example extracts the Json from a log file.