REGEXP Match * through multiple lines ?

ALInux · 08-13-2007, 04:00 AM

Hi All
I am learning about regexp right now, I read at
http://www.regular-expressions.info/dot.html the following

The dot matches a single character, without caring what that character is. The only exception are newline characters. In all regex flavors discussed in this tutorial, the dot will not match a newline character by default. So by default, the dot is short for the negated character class [^\n] (UNIX regex flavors) or [^\r\n] (Windows regex flavors).

The question is how to match any character through multiple lines. I.e. START..multiple lines .* END

Thx for any help

wjevans_7d1@yahoo.co · 08-13-2007, 07:56 AM

Wherever you use

Code:

replace it with

Code:

[.\n]

Hope this helps.

raskin · 08-13-2007, 08:06 AM

It doesn't work. [.] is simply dot. I'd bet on

Code:

(.|[\n])

It works in Vim (in verymagic mode, you'll need

Code:

\(.\|[\n]\)

usually). But in Vim you can just use

Code:

\_.

.

ALInux · 08-13-2007, 08:07 AM

To: wjevans_7d1@yahoo.co
The thing is that the search words are not located at the beginning of the line..so it does not work..I need to match at any location not just at the start of the line...I mean that is how I understood .\n
I tried it ..it did not work

Thx for the help so far any more suggestions ?

I will try raskin's suggestion now

ALInux · 08-13-2007, 08:11 AM

[root@localhost logs]# cat call.log | sed -n -e '/IN.|[\n]From: 9613041705 <sip/','/Contact: "9613041705"/p'

Did not return any results.

AAA = IN
BBB = From: 9613041705 <sip
CCC = Contact: "9613041705

I need to get the text from AAA to CCC with BBB included , the text spawns over multiple lines..and the location of the words differs.
Any suggestions please..

ALInux · 08-13-2007, 08:24 AM

[root@localhost logs]# cat call.log | sed -n -e '/IN(.*|.\n)From: 9613041705 <sip/','/Contact: "9613041705"/p'

This did not work either.

raskin · 08-13-2007, 08:31 AM

Sed doesn't play nice with newlines anyway. If you find how to convert all newlines with to pluses, I will modify it to do what you want from it now.. But I think it will not wok well.

ALInux · 08-13-2007, 10:32 AM

Thank you that would be great...can u suggest any other way to do it if sed does not play along well

spirit receiver · 08-13-2007, 10:49 AM

Maybe something like this?

Code:

sed -ne '/AAA/,/CCC/p' filename

theYinYeti · 08-13-2007, 10:54 AM

sed and awk won't do that. perl and php can though

Yves.

spirit receiver · 08-13-2007, 11:01 AM

Seems like I'm the only one who doesn't understand the question...

ALInux · 08-14-2007, 01:51 AM

Dear spirit receiver your code does work well however it does not take into account that BBB must be found in between those two entities..I.e. it matches

AAA-----no BBB------>CCC Wrong
&&
AAA------BBB-------->CCC Right..

Can you help me to adjust the code accordingly ?
Thx

spirit receiver · 08-14-2007, 07:39 AM

This might work (but probably not with all flavors of sed):

Code:

sed -ne '/AAA/{:LOOP /CCC/{b CHECK};N;b LOOP;:CHECK /BBB/p}' filename

When /AAA/ matches, all lines until /CCC/ are added to the pattern space. Then the pattern space is checked for /BBB/.