GNU bug report logs - #21865
Parenthesis subexpressions

Previous Next

To reply to this bug, email your comments to 21865 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Valerio Bozzolan <bozzolan.valerio <at> educ.di.unito.it>
To: bug-grep <at> gnu.org
Subject: Parenthesis subexpressions
Date: Sun, 08 Nov 2015 21:42:44 +0100

[Message part 1 (text/plain, inline)]

Hi,

(First time in a GNU mailing list!)

I've already asked this question to my local GNU/Linux user group and in #grep <at> Freenode... I'm still confused.

GNU Grep don't have an arg to choose the subexpression. Right?

Stupid e.g.:
    echo abcde | grep -o -E 'b([a-z])d'
    => "bcd"

What if I want the first subexpression? ("b")? GNU Grep can't do it. Isn't it? (Why?)

I actually use GNU Awk, or GNU Bash with $BASH_REMATCH[$n_sub].

Thank you for the clarification!
--
Valerio Bozzolan
Email sent from Android (CyanogenMod) using K-9 Mail.

[Message part 2 (text/html, inline)]

Message #8 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Valerio Bozzolan <bozzolan.valerio <at> educ.di.unito.it>
To: bug-grep <at> gnu.org
Subject: Re: Parenthesis subexpressions
Date: Sun, 08 Nov 2015 21:49:03 +0100

[Message part 1 (text/plain, inline)]

Sorry... typo...

    echo abcde | grep -o -E 'b([a-z])d'
    => "bcd"

Can't I choose to have only "c"?

Thanks again!

On 8 November 2015 21:42:44 CET, Valerio Bozzolan <bozzolan.valerio <at> educ.di.unito.it> wrote:
>Hi,
>
>(First time in a GNU mailing list!)
>
>I've already asked this question to my local GNU/Linux user group and
>in #grep <at> Freenode... I'm still confused.
>
>GNU Grep don't have an arg to choose the subexpression. Right?
>
>Stupid e.g.:
>    echo abcde | grep -o -E 'b([a-z])d'
>    => "bcd"
>
>What if I want the first subexpression? ("b")? GNU Grep can't do it.
>Isn't it? (Why?)
>
>I actually use GNU Awk, or GNU Bash with $BASH_REMATCH[$n_sub].
>
>Thank you for the clarification!
>--
>Valerio Bozzolan
>Email sent from Android (CyanogenMod) using K-9 Mail.

-- 
Valerio Bozzolan
Email sent from Android (CyanogenMod) using K-9 Mail.

http://boz.reyboz.it

[Message part 2 (text/html, inline)]

Message #11 received at 21865 <at> debbugs.gnu.org (full text, mbox):

From: Stephane Chazelas <stephane.chazelas <at> gmail.com>
To: Valerio Bozzolan <bozzolan.valerio <at> educ.di.unito.it>
Cc: 21865 <at> debbugs.gnu.org
Subject: Re: bug#21865: Parenthesis subexpressions
Date: Mon, 9 Nov 2015 13:50:46 +0000

2015-11-08 21:49:03 +0100, Valerio Bozzolan:
> Sorry... typo...
> 
>     echo abcde | grep -o -E 'b([a-z])d'
>     => "bcd"
> 
> Can't I choose to have only "c"?
[...]

That's correct, GNU grep doesn't have that capability (yet).
Recent versions of pcregrep do:

$ echo abc | pcregrep -o1 '.(.).'
b

Now, I'm not a GNU grep maintainer but I suppose the question is
how far do we want to take grep away from its original purpose
(print the lines that match a pattern which is what g/re/p
stands for).

GNU grep is already doing find's job with -r, part of sed's job
with -o/--colour.

Having said that, I do agree it's the logical continuation after
-o.

Note that for now, you can already do:

$ echo abcde | grep -o -P 'b\K[a-z](?=d)'
c


-- 
Stephane

Message #14 received at 21865 <at> debbugs.gnu.org (full text, mbox):

From: Valerio Bozzolan <bozzolan.valerio <at> educ.di.unito.it>
Cc: 21865 <at> debbugs.gnu.org
Subject: Re: bug#21865: Parenthesis subexpressions
Date: Mon, 09 Nov 2015 17:03:39 +0100

[Message part 1 (text/plain, inline)]

Thanks for agreeing with the evolution of the meaning of "-o".

Just to make you a laugh: I was reproducing egrep with $BASH_REMATCH:
https://gist.github.com/valerio-bozzolan/6787675e931dce1ba7e9

Definitely not beautiful... but really effective for me.

So something like "egrep -o $n regex" also can save the world from code similar to mine.

On 9 November 2015 14:50:46 CET, Stephane Chazelas <stephane.chazelas <at> gmail.com> wrote:
>2015-11-08 21:49:03 +0100, Valerio Bozzolan:
>> Sorry... typo...
>> 
>>     echo abcde | grep -o -E 'b([a-z])d'
>>     => "bcd"
>> 
>> Can't I choose to have only "c"?
>[...]
>
>That's correct, GNU grep doesn't have that capability (yet).
>Recent versions of pcregrep do:
>
>$ echo abc | pcregrep -o1 '.(.).'
>b
>
>Now, I'm not a GNU grep maintainer but I suppose the question is
>how far do we want to take grep away from its original purpose
>(print the lines that match a pattern which is what g/re/p
>stands for).
>
>GNU grep is already doing find's job with -r, part of sed's job
>with -o/--colour.
>
>Having said that, I do agree it's the logical continuation after
>-o.
>
>Note that for now, you can already do:
>
>$ echo abcde | grep -o -P 'b\K[a-z](?=d)'
>c
>
>
>-- 
>Stephane

-- 
Valerio Bozzolan
Email sent from Android (CyanogenMod) using K-9 Mail.

[Message part 2 (text/html, inline)]

Message #17 received at 21865 <at> debbugs.gnu.org (full text, mbox):

From: Stephane Chazelas <stephane.chazelas <at> gmail.com>
To: Valerio Bozzolan <bozzolan.valerio <at> educ.di.unito.it>
Subject: Re: bug#21865: Parenthesis subexpressions
Date: Mon, 9 Nov 2015 17:16:19 +0000

2015-11-09 17:00:36 +0100, Valerio Bozzolan:
> Thanks for agreeing with the evolution of the meaning of "-o".
> 
> Just to make you a laugh: I was reproducing egrep with $BASH_REMATCH:
> https://gist.github.com/valerio-bozzolan/6787675e931dce1ba7e9
> 
> Definitely not beautiful... but really effective for me.

You may want to read:

https://unix.stackexchange.com/questions/169716/why-is-using-a-shell-loop-to-process-text-considered-bad-practice
https://unix.stackexchange.com/questions/209123/understand-ifs-read-r-line
https://unix.stackexchange.com/questions/65803/why-is-printf-better-than-echo

Here, if there wasn't a pcregrep already, I'd rather do it in
perl or GNU sed than bash.

Like:

perl -lne 'print for /a([a-z])c/g'

Also note that:

echo abac | pcregrep -o1 'a(.)'
b
c

> So something like "egrep -o $n regex" also can save the world from code similar to mine.

GNU grep can't add it like that as that would break backward
compatibility.

grep -o 1 regex file

is currently meant to print the occurrences of "1" in the
"regex" and "file" files.

Even adding it as:

grep -o1 regex file

would probably not be a good idea as that would mean some
ad-hoc parsing of the options (in "grep -o1 regexp", "1" would
be an argument to "-o" while in "grep -oi regexp", "i" currently
is a separate "-i" option).

So reasonably, it should probably be a separate option like -O 1.

-- 
Stephane

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.