Your Unix/Linux Chapter 12

Dr. Deborah Whitfield

Chapter 12: awk

(see man awk!)
awk options 'selection_criteria {action}' file
awk [-F regexpr] -f progfile [-v asgn] [ arg ]
-F sets the field separator to the regular expression
-f Read the awk program from the progfile rather than the command line
-v Perform the awk assignment asgn before executing the awk program

Samples
awk '/browns/ { print }' foo

Lines containing browns
awk '$2 ~ /^browns$/ { print }' foo

Field 2 is browns
awk 'NR == 1, NR ==5 { print }' foo

Records 1 to 5 inclusive
awk '$6 > 2000 { print }' foo

Field 6 is greater than 2000

An awk program consist of repeated statements that can be preceded by initializations and can be ended with final actions. Actions are enclosed in { }. The awk commands are placed in single quotes like sed commands.

BEGIN { action } - perform the actions before processing the file
END {action} - perform the actions after processing
{ action } - perform actions to each input record
pattern [ {action} ] - perform action on only those records that match the pattern
pat1,pat2[{action}] - perform action in the range of records from pat1 to pat2

Rules for statements in an awk program:

Pattern and its opening left brace { must appear on the same line
BEGIN {and END { must be on the same line
Separate statements with newlines or ;
To continue statements onto the next line, end all but the last line with a \
Comments start with # to end of line
Blank lines are ignored
Spaces and tabs can be used around awk elements, but are not required unless the two elements together could be 1 element.
A left parenthesis when used for function or a statement such as print or close, must be adjacent to the function or statement name (NO intervening whitespace).

Awk elements consist of constants (numbers and strings), variables, regular expressions, array elements, function calls, and parenthesized expressions.
Unlike the shell, mathematical operations do not need a call to expr.
User defined variables mimic the shell; however $0, $1, ... take on a different meaning.
Awk separates the entire input record into fields separated by whitespace. $1 is the first field, $2 is the second and $0 is the entire record.
The predefined variable FS. Assigning FS a different value (assignment statement or the -F flag), will cause awk to use that value as a separator.
Similarly, RS is the record separator and can be set to a single character. The default is \n.

Awk Operators

= += -+ *= /= %=
Assignment just like C
|| &&
Or And
!
Not
> >= < <= == !=
Relational Operators (numerical)
~? !~
Match and Non-Match (string)
+ - * / %
Typical math operators
++ --
Increment and Decrement

Control Structures

If statement is just like C's
if ( condition )
Stmt1
else
Stmt2

Built-In Variables (Section 12.10)

NR - number of records
FS - field separator
OFS - output field separator
RS - record separator
NF - number of fields in current line
FILENAME - current input file
ARGC - command line argument count
ARGV - array containing list of arguments

For statement is like C's NOT the shells

for (exp1; cond; exp2)
Stmt

While Statement (equivalent to above for):
exp1

while ( condition) {
stmt
exp2
}

awk Built-in functions (Section 12.11)

cos(expr)
Cosine of expr
exp(expr)
Exponential of expr : e**expr
getline()
Read next line; 0 if EOF, 1 if not
index(s1, s2)
Position of string s2 in s1; returns 0 if not present
int(expr)
Integer part of expression, truncates
length(s)
Length of string s
Log(expr)
Natural logarithm of expr
Sin(expr)
Sine of expr
split(s, a, c)
Splits s into a[1], ... a[n] on character c; returns n
sprintf(fmt, ...
Format according to fmt (like C's printf)
substr(s, m)
substring of s beginning at position m
substr(s, m, n)
n-character substring of s beginning at position m
tolower(s)
returns lowercase value of string s
toupper(s)
returns uppercase value of string s

awk examples

All work herein is subject to copyright. Original content to Dr. Deborah Whitfield, text content (Your UNIX/Linux) to Prentice Hall publishing.