awk: What pattern has the same meaning with egrep -w ?

quanba · 10-06-2009, 03:29 AM

Hi there,

when I was processing text, I got the problem:

1: test
2: adsf;test;dfas
3: asdf test asdfa
4: asdf-test-bcd
5: adb,test,add
6: atestabc

I can grep the word "test" with egrep -w, but I want to process the text with awk, what pattern do I need?

thanks in advance.

neonsignal · 10-06-2009, 06:19 AM

From 'man awk':

Quote:

\y matches the empty string at either the beginning or the end of a word.
\B matches the empty string within a word.
\< matches the empty string at the beginning of a word.
\> matches the empty string at the end of a word.
\w matches any word-constituent character (letter, digit, or underscore).
\W matches any character that is not word-constituent.

So to match the word test, you would use:

Code:

awk '/\<test\>/'

pixellany · 10-06-2009, 06:28 AM

Quote:

Originally Posted by quanba

I can grep the word "test" with egrep -w, but I want to process the text with awk,

It would help to say what you actually need to do, and why you need to use AWK. For simple search and replace, I would use SED.

The "pattern" you refer to is typically called a "Regular Expression" (Regex), and is using in many utilities.

pixellany · 10-06-2009, 06:40 AM

Quote:

\y matches the empty string at either the beginning or the end of a word.
\B matches the empty string within a word.
\< matches the empty string at the beginning of a word.
\> matches the empty string at the end of a word.
\w matches any word-constituent character (letter, digit, or underscore).
\W matches any character that is not word-constituent.

Interesting---I had never seen some of these. They all work in SED except "\y"......

Tinkster · 10-06-2009, 03:04 PM

Quote:

Originally Posted by quanba

Hi there,

when I was processing text, I got the problem:

1: test
2: adsf;test;dfas
3: asdf test asdfa
4: asdf-test-bcd
5: adb,test,add
6: atestabc

I can grep the word "test" with egrep -w, but I want to process the text with awk, what pattern do I need?

thanks in advance.

From that request it's not quite clear what you're
after. If you just need to find lines that match the
word test, then /<test>/ will print those lines.

If you want to operate on the matched string alone,
that's not going to help you.

You'd be looking at something somewhat more complex.

Code:

'{res=gensub(/.*(\<test\>).*/, "\\1","g"); if(res!=$0){print res}}'

Btw, line 6 in your sample won't be matched by either the
egrep or the awk ...

quanba · 10-11-2009, 11:35 PM

Thanks for your supports. But I was missing information from the title. I am practicing on HPUX and none of the patterns works.

Have anyone experience the problem on HPUX?

pixellany · 10-12-2009, 12:22 AM

Quote:

Originally Posted by quanba

Have anyone experience the problem on HPUX?

Have you checked the man pages and other documentation?

I'm moving this thread since it is not a Linux question......good luck.