AkiRoss' AWK cheatsheet
by Alessandro [AkiRoss] Re ~ www.ale-re.net
-------------------------------------------------------------------------------

**Big helps:
http://www.vectorsite.net/tsawk.html
http://www.gnu.org/manual/gawk/html_node/index.html

**My help (a cheatsheet):
AWK stands for Aho, Weinberger, Kernighan.
Quick console usage
	awk -f script_file input_data
	awk 'pattern { action }' input_data
	awk 'rules' -			Read input from stdin
Script are made of rules:
	pattern { action }
	action is executed if pattern is matched
	you can omit pattern to execute that action on every row
	you can omit action to print the row that match your pattern
Syntax is really similar to C one
	{ block }
	if (cond) ... else ...
	# Comments
	while, for, do, switch, break, continue, next, exit
BEGIN and END patterns are fixed and executed when script begin and end
Usual operators
	= (assign)
	== (compare)
	++, +=, ...
Strings are "as usual", but AWK can evaluate them as integer
	"123" == 123
In addition there are regexps, quoted between /s
	/^[Rr]eg[Ee]xp?$/
	metachars doesn't need quoting
	usual metachars
Patterns are automagically divided into fields
	Fields are separated by a given char (default is ' ') or regex
	$0						Is the whole line
	$n						Is the n-th field (n > 0)
	Field variables aren't read only, you can write into them
	Modify the field separator with -F on the command line or use the FS var
	FS						Can be a char or a regex.
							In this case, regex is quoted with "" instead of //
Patterns are evaluated expressions:
	/regex/					matches lines with contains regex
							(e.g. regex, regexASD, ASDregexASD)
	/^regex$/ 				to match exactly that regex
	"matchall"				matches all lines
	""						matches no line
	"123" == 123 			is a valid pattern
	($1 == "hello") &&		boolean composition
	($2 == "folk")
There is the ~ operator for matching regexp
	$1 ~ /regex/			first field match regex
	$2 !~ /regex/			second field doesn't match regex
You can use variables in your script
	They don't need to be declared
	Default value is 0 or "" for unitialized variables
	Auto conversion between math and strings
	There are pre-defined variables
		Some of them are:
		NR					Keeps how many rows in input has been processed
		FS					Field separator variable, can be char or regex (default spaces)
		OFS					Output field separator (default spaces)
		NF					Number of fields on this line
		FILENAME			Guess...
You can use some functions for your actions
	print is the default action. Does newline.
		print 				behaves like print $0
		print param			print the parameters
		print a,b,c 		print a, b and c separed by OFS
	printf formatted print. Works like in C
		printf("Your output, Mr %s\n", "John");
	sprintf return a string with printf output
	sqrt(), log(), exp(), int()
	substr(), split(), index()
Redirections and pipes
	print > "file"			Print on file (create and overwrite)
	print >> "file"			Print on file (create and append)
	print | "tool opts"		Pipes the output to a shell tool
	they works on outputs, not variables or return values

Should be enough :)
Stay --sync