Need help with grep and regex

MBoyle19 · 04-10-2017, 01:26 AM

Quote:

Originally Posted by pan64

glad to help you
if you really want to say thanks just click on yes

And this is the case when you do not need (but in general yes, you have to use quotes).

grep knows 3 different kind of regex, sed knows at least 2, awk has its own style, but more or less similar....
perl has its own PCRE and python knows that too.

Just clicked on yes

Quote:

Originally Posted by Turbocapitalist

syg00 mentioned them above in #2 and #4. They are perl regular expressions, ported to other languages this is often called perl-compatible regular expression (PCRE). Within perl they are properly called metasymbols and they stand in for some functionality as well as pattern data. You'll find PCRE nearly everywhere these days.

But as to whether sed and awk can understand perl's regular expression metasymbols, the answer is, "no" at least for existing versions to-date.

If you find them useful, then you can get a lot out of one-liners in perl. See the options -e, -i, -n, and -p in

Code:

man perlrun

Anything you can do with sed and awk you can do, plus more, in perl.

Regex is complex. I read online on it's usage and some stuff is still Greek to me. To learn another language like perl is going to give me a brain meltdown. Maybe I give it a quick look some day

Turbocapitalist · 04-10-2017, 01:33 AM

You don't have to learn it all at once, just like you don't have to learn the languages sed and awk all at once either. Here you can use it like beginning-level sed:

Code:

perl -p -e 's/\w+/blah/;' < inputfile.txt > outputfile.txt

perl -p -i.bak -e 's/\w+/blah/;' *.txt

Those formulas will allow you to do basic sed-like substitution but with full perl regular expressions including your full range of metasymbols and special variables.

MBoyle19 · 04-10-2017, 01:34 AM

Thank you all! I clicked on yes for helpful answers to ALL members here

pan64 · 04-10-2017, 01:47 AM

and actually there are online regex testers available: http://www.myezapp.com/apps/dev/regexp/show.ws or http://www.regextester.com/

MBoyle19 · 04-10-2017, 01:54 AM

Quote:

Originally Posted by Turbocapitalist

You don't have to learn it all at once, just like you don't have to learn the languages sed and awk all at once either. Here you can use it like beginning-level sed:

Code:

perl -p -e 's/\w+/blah/;' < inputfile.txt > outputfile.txt

perl -p -i.bak -e 's/\w+/blah/;' *.txt

Those formulas will allow you to do basic sed-like substitution but with full perl regular expressions including your full range of metasymbols and special variables.

Quote:

Originally Posted by pan64

and actually there are online regex testers available: http://www.myezapp.com/apps/dev/regexp/show.ws or http://www.regextester.com/

and a

TobyV · 04-10-2017, 05:54 AM

Quote:

Originally Posted by MBoyle19

Can sed and awk understand shorthand regex?

sed can use shorthand or extended regex with the -r switch. Don't know about awk.

For example, I want to pad a zero to the number 1 and not 3 using the \b word boundary.

Code:

echo track 1 of 3 | sed -r 's/\b[[:digit:]]\b/0&/'
track 01 of 3

Below is the gnu-style regex for word boundary

Code:

echo track 1 of 3 | sed 's/\<[[:digit:]]\>/0&/'
track 01 of 3

The [:digit:] is called a class set, It is equivalent to [0-9]