Customer Reviews

Effective awk Programming: Text Processing and Pattern Matching

5 star:		(2)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

Awk is a very wonderful language. Usually, we like heap praise on languages that let us do complex things in simple, human-readable and easily understandable ways. For example, you could say that Java is better than C for internet connectivity in this regard, since in Java, you just need to do is call socket(IP_address, port), rather than mess around with calls to a whole library of functions.

But Awk is brilliant in an entirely different way, in a UNIX way. Want to know how many directories you have in your tree? just enter
"ls -l -R | awk '$1~/^d/ {s++} END {print s}'" (this might take a while if you have a lot in your directory!) Done. The brilliance of awk is in its simplicity, but the primary obstacle to most is its syntax. You really need to know about UNIX regular expressions, the way that other UNIX utilities behave, etc. These are things that smart people don't memorize, but keep a good reference book handy, and that is why you NEED this book.

Sooner or later, if you use UNIX, you're going to need to use awk. Chances are, actually, that if you already use UNIX and you don't know awk, you've already wasted countless hours writing little utilities in languages like C to process text files - utilities that could have been implemented with a few lines of awk. The fact that you can experiment with awk on the command line, and then put what you've learned into an awk script is one of the great strengths of the language.

Arnold Robbins has done a fantastic job in this book of guiding the reader through the dark and sometimes scary, but always useful awk language. This book, actually, is the official manual for gawk, the GNU implementation of awk. Oh, you didn't know that there were different implementations of awk, and they don't all agree about some of the finer nuances of the language? Then you REALLY need this book, because these variations are all discussed as they come up (not hidden in some appendix at the back!).

When I "discovered" awk for myself, I looked through a lot of guides, and this one won hands down. I have my pile o' computer references, but this one sits on the shelf right where I work.

...and one more thing, "awk" isn't short for awkward, it's the initials of the three original writers of the language!

Awk is one of those handy Unix tools with which you can easily impress people. Using a simple

/pattern/ { action }

syntax, you can construct powerful one-liners. Do you want to how much time in total you spent surfing the Internet? Here it is:

awk '/Connect time/ { s += $8 } END { print s }' /var/log/messages

It doesn't get much shorter in any other programming language, does it? Need to strip text of HTML tags? Need a frequency count of words in a text? Awk is the perfect tool for tasks like this. With its pattern-action structure, powerful regular expression mechanism, associative arrays and basic program flow control, it provides a powerful tool for manipulating flat text files. Even though other scripting languages may be richer in features, there exists a niche where Awk is just the right tool to do the job.

Arnold Robbins, the author of this book as well as of several other books on Awk, serves also as the maintainer of GNU Awk (gawk for short), the most influential version of Awk available today. With the version 3.10, released in 2001, GNU Awk became richer for a handful of new extensions over traditional Awk, most important among them are the TCP/IP networking and the support for internationalization. All new extensions are described in the book. How successful these new extensions will be is doubtful, however. Networking scripting niche is already well covered with Perl and Python, and internationalization doesn't really matter much in short throw-away scripts Awk is usually used for.

With all due respect to the creators of Awk and their book (Alfred V. Aho, Peter J. Weinberger, Brian W. Kernighan, The Awk Programming Language, Addison-Wesley, 1988), I have to say that "Effective Awk Programming" is probably the best Awk tutorial on the market today. If you are serious about learning Awk, you shouldn't be without it. If you are still hesitating whether it wouldn't be wiser investing those 28 USD elsewhere, here is chance to read it before you buy it: install GNU Awk 3.10, and the Texinfo source of the book comes with it. But sooner or later you will find O'Reilly RepKover binding too tempting...