Thursday, July 12, 2007

AWK: a short introduction

HomeMapIndexSearch NewsArchivesLinksAbout LF
[Top bar]
[Bottom bar]
[Photo of the Author]
by Javier Palacios Bermejo

About the author:

Javier is involved in a Ph. D. in Astronomy at a Spanish university where he administrates a workstation cluster. The daily work in his department is done on Unix machines. After some initial problems and trials slackware Linux was chosen. Linux turned out to be much better than some other proprietary Unix systems.


Examples with awk: A short introduction



This article gives some insight in to the tricks that you can do with AWK. It is not a tutorial but it provides real live examples to use.

Originally, the idea to write this text came to me after reading a couple of articles published in LinuxFocus that were written by Guido Socher. One of them, about find and related commands, showed me that I was not the only one who used the command line. Pretty GUIs don't tell you how the things are really done (that's the way that Windows went years ago). The other article was about regular expressions. Although regular expressions are only slightly touched in this article, you need to know them to get the maximum from awk and other commands like sed and grep.

The key question is whether this awk command is really useful. The answer is definitly yes! It could be useful for a normal user to process text files, re-format them etc... For a system administrator AWK is really a very important utility. Just walk around /var/yp/Makefile or look at the initialization scripts . AWK is used everywhere.  

Introduction to awk

My first news about AWK are old enough for being forgotten. I had a colleague who needed to work with some really big outputs from a small Cray. The manual page for awk on the Cray was small, but he said that AWK looks very much like the thing he needs although he did not yet understand how to use it.
A long time later, we are back in my life again. A colleague of mine used AWK to extract the first column from a file with the command:

awk '  '{print $1}'   file

Easy, isn't it? This simple task does not need complex programming in C. One line of AWK does it.

Once we have learned the lesson on how to extract a column we can do things such as renaming files (append .new to "files_list"):

ls files_list | awk '{print "mv "$1" "$1".new"}' | sh

... and more:

  1. Renaming within the name:
    ls -1 *old* | awk '{print "mv "$1" "$1}' | sed s/old/new/2 | sh
    (although in some cases it will fail, as in file_old_and_old)

  2. remove only files:
    ls -l * | grep -v drwx | awk '{print "rm "$9}' | sh
    or with awk alone:
    ls -l|awk '$1!~/^drwx/{print $9}'|xargs rm
    Be careful when trying this out in your home directory. We remove files!

  3. remove only directories
    ls -l | grep '^d' | awk '{print "rm -r "$9}' | sh
    ls -p | grep /$ | wk '{print "rm -r "$1}'
    or with awk alone:
    ls -l|awk '$1~/^d.*x/{print $9}'|xargs rm -r
    Be careful when trying this out in your home directory. We remove things!

  4. killing processes by name (in this example we kill the process called netscape):
    kill `ps auxww | grep netscape | egrep -v grep | awk '{print $2}'`
    or with awk alone:
    ps auxww | awk '$0~/netscape/&&$0!~/awk/{print $2}' |xargs kill
    It has to be adjusted to fit the ps command on whatever unix system you are on. Basically it is: "If the process is called netscape and it is not called 'grep netscape' (or awk) then print the pid"

As you can see, AWK really helps when the same calculations are repeated over and over ... and apart from that it is much more fun to write an AWK program than doing almost the same thing 20 times manually.

awk is a little programming language, with a syntax close to C in many aspects. It is an interpreted language and the awk interpreter processes the instructions.

About the syntax of the awk command interpreter itself:

# gawk --help
Usage: gawk [POSIX or GNU style options] -f progfile [--] file ...
gawk [POSIX or GNU style options] [--] 'program' file ...
POSIX options: GNU long options:
-f progfile --file=progfile
-F fs --field-separator=fs
-v var=val --assign=var=val
-m[fr] val
-W compat --compat
-W copyleft --copyleft
-W copyright --copyright
-W help --help
-W lint --lint
-W lint-old --lint-old
-W posix --posix
-W re-interval --re-interval
-W source=program-text --source=program-text
-W traditional --traditional
-W usage --usage
-W version --version

Instead of simply quoting (') the programs in the command line, we can, as you can see above, write the instructions into a file, and call it with the option -f. With command line defined variables using -v var=val we can add some flexibility to the programs.

Awk is, roughly speaking, a language oriented to manage tables. That is some information which can be grouped inside fields and records. The advantage here is that the record definition (and the field definition) is flexible.

Awk is powerful. It's designed for work with one-line records, but that point could be relaxed. In order to see in some of these aspects, we are going to look at some illustrative (and real) examples.

  • Printing tables in a slightly prettier way
    Maybe, you have had to print some ASCII table obtained from somewhere. For example the hostnames, ethernet and IP numbers in a list. When those tables are really big, reading becames difficult, and we feel that we need this list printed with LaTeX or, at least, with a better format. If the table is simple then it's not too dificult:

    BEGIN {
    printf "LaTeX preample"
    printf "\\begin{tabular}{|c|c|...|c|}"
    { printf $1" & "
    printf $2" & "
    printf $n" \\\\ "
    printf "\\hline"
    END {
    print "\\end{document}"

    Certainly, this is not a generic program, but we're just starting ...
    (The double backslashes (\) are necessary because it's the shell escape character)

  • Slicing output files
    SIMBAD is an astronomical objects database that, among other things, provides a stars positions on the sky plane. Once in the past I needed to perform searches to draw charts around some objects. The interface allowed to save the results in text files, and I had two approaches: 1) create one file for each object, or 2) feed it with the whole input list, getting a single big output log file with the query results. As I decided to go for the second approach, I used awk for slicing the big output log. Obviously, I needed to take advantage on some output characteristics.

    1. Each request produces a header line with a format like
      ====> name : nlines <====
      The first header allow us to know when a new object begans, and the fourth how many entries the object contains.

    2. The character used in the output lists to mark different columns was '|'. This requires two additional code lines to filter to the output and get only the fields I was interessted in.
    ( $1 == "====>" ) {
    NomObj = $2
    TotObj = $4
    if ( TotObj > 0 ) {
    FS = "|"
    for ( cont=0 ; cont<TotObj ; cont++ ) {
    print $2 $4 $5 $3 >> NomObj
    FS = " "
    Acutally, the object name was not returned, and it was sligthly more complicated, but this is supposed to be an illustrative example.

  • Playing with the mail spool

    BEGIN {
    BEGIN_MSG = "From"
    BEGIN_BDY = "Precedence:"
    MAIN_KEY = "Subject:"

    HEAD = "NO"; BODY = "NO"; PRINT="NO"
    OUT_FILE = "Month_Reports"


    if ( $1 == BEGIN_MSG ) {
    HEAD = "YES"; BODY = "NO"; PRINT="NO"

    if ( $1 == MAIN_KEY ) {
    if ( $2 == VALIDATION ) {
    PRINT = "YES"
    $1 = ""; $2 = ""
    print "\n\n"$0"\n" > OUT_FILE

    if ( $1 == BEGIN_BDY ) {
    if ( $0 == "" ) {
    HEAD = "NO"; BODY = "YES"
    } else {
    HEAD = "NO"; BODY = "NO"; PRINT="NO"

    if ( BODY == "YES" && PRINT == "YES" ) {
    print $0 >> OUT_FILE
    Maybe we are administrating a mailing list and from time to time, some special messages are submitted to the list (for example, monthly reports) with some specific format (subject as '[MONTH REPORT] month , dept'). Suddenly, we decide at the end of the year put together all these messages, saving aside the others.
    This can be done by processing the mail spool with the awk program on the left.

    To get each report written to an individual file means three extra lines of code.

    NOTE: This example assumes that the mail spool is structured as I think it is. This programs works for my mail.

I've used awk for many other tasks (automatic generation of web pages with information from simple databases) and I know enough about awk programming to be sure that a lot of things can be done.
Just let your imagination fly.


A problem

One problem is that awk needs perfect tabular information, no holes, awk does e.g not work with fixed width columns. This is not problematic if we create by ourself the awk input: choose something uncommon to separate the fields, later we fix it with FS and we are done!!! If we already have the input this could be a little more problematic. For example a table like this:
1234  HD 13324  22:40:54 ....
1235 HD122235 22:43:12 ....
This is difficult to handle this with awk. Unfortunately this is quite common. If we have only one column with this characteristics, we can solve the problem (if anybody knows how to manage more than one column in a generic case, please let me know!).
I had to face one of these tables, similar to the one described above. The second column was a name and it included a variable number of spaces. As it usually happens, I had to sort it using the last column.

... and a solution

I realized that the column I wanted to sort was the last one and awk knows how many fields there are in the current registry. Therefore, it was enough to access the last one (sometimes $4, and sometimes $5, but always NF). At the end of the day, the desired result was obtained:
awk '{ printf $NF;$NF = "" ;printf " "$0"\n" }' | sort

This just shifts the last colum to the first position and you can sort it. Obviously, this method is easily applied to the third field starting from the end, or to the field which goes after a control field which has always the same value.
Just use your ideas and imagination.


Deeper AWK


Working over matched lines

Up to now, nearly all the examples process all the input file lines. But, as also the manual page states, it is possible to process only some of the input lines. One must just preceed the group of commands with the condition the line should meet. The matching condition could be very flexible, variing from a simple regular expression to a check on the contents of some field, with the possibility of grouping conditions with the proper logical operators.


Awk as a programming language

As any other programming language, awk implements all the necessary flow control structures, as well as a set of operators and predefined functions to deal with numbers and strings.

It's possible, of course, to include user defined functions with the keyword function. Apart from the common scalar variables, awk is also able to manage variable sized arrays.


Including libraries

As it happens in any programming language, there are some very common functions and it becomes uncomfortable to cut and paste pieces of code. That's the reason why libraries exist. With the GNU version of awk, is possible include them within the awk program. This is however an outlook to the things which are possible and outside the scope of this article.



Certainly, awk might not be as poweful as many other tools designed with similar goals. But it has the big advantage that it is possible write in a really short time small programs which are fully tailored to our needs.

AWK is very appropriate for the purposes for which it was build: Read data line by line and act upon the strings and patterns in the lines.

Files like /etc/password turn out to be ideal for reformatting and processing with AWK. AWK is invaluable for such tasks.

Of course AWK is not alone. Perl is a strong competitor but still it is worthwhile to know some AWK tricks.


Additional information

This kind of very basic commands and is not very well documented, but you can find something when looking around.

  • awk syntax is not the same in every Unix system, but there is a way to learning how is it in our particular system:
    man awk

  • O'Reilly has published a book: Sed & Awk (Nutshell handbook) by Dale Dougherty.

  • Looking at Amazon, we find more titles such as Effective Awk Programming: A User's Guide, oriented to gawk, and half a dozen titles more.

Usually, all books on unix mention this command, but only some of them treat it in detail. The best we can do, is to browse any book we get into our hands. You never know where useful information can be found.

Webpages maintained by the LinuxFocus Editor team
© Javier Palacios Bermejo
LinuxFocus 1999
Translation information:

es->--Javier Palacios Bermejo
es->enJavier Palacios Bermejo,Ruben Sotillo, Manuel Rodriguez

1999-10-14, generated by lfparser version 0.7