topic egrep and patterns in Operating System - Linux

egrep and patterns

Leo The Cat — Fri, 05 Jun 2009 17:13:28 GMT

Hi

I know to use egrep like this for the OR
egrep -i '(65535|65536)'

How to do the same thing but with AND

Any Idea

Bests Regards
Den

Re: egrep and patterns

Leo The Cat — Fri, 05 Jun 2009 17:24:47 GMT

Ho sorry

precision: patterns are not in the same line !

Re: egrep and patterns

Leo The Cat — Fri, 05 Jun 2009 17:25:50 GMT

Here my input
a
b
c
d
e
test1
f
g
f
test2
h
h
j

I need to find if this input contains test1 AND test2 !!!

Re: egrep and patterns

Ivan Ferreira — Fri, 05 Jun 2009 18:18:32 GMT

Please refer to this thread:

http://forums13.itrc.hp.com/service/forums/questionanswer.do?admit=109447627+1244229489824+28353475&threadId=1320198

Re: egrep and patterns

James R. Ferguson — Fri, 05 Jun 2009 20:19:47 GMT

Hi Den:

This is one way (if you insist on 'grep') but requires two passes of your input:

# grep -q test1 myfile && grep -q test2 myfile && echo "ok" || echo "no match"

A faster way would be to use Perl or awk like:

# perl -nle 'm/test2/ and $x=1;m/test1/ and $y=1;last if $x+$y==2;END{print $x+$y==2 ? "ok" : "nomatch"}' myfile

By the way, you have un-evaluated solutions in your earlier thread:

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1344958

Feedback there too is appreciated :-)

Regards!

...JRF...

Re: egrep and patterns

James R. Ferguson — Fri, 05 Jun 2009 23:29:33 GMT

Hi (again) Den:

In the event that you want more than a "yes/no" answer to the presence of the AND'ed match (as I provided in the first solutions, above) you could use:

# grep -q test1 myfile && grep -q test2 myfile && grep -E "test1|test2" myfile

...which again "suffers" from multiple passes through the file.

...OR in one pass of the file:

# perl -nle 'm/test1/ and $x=1,push(@a,$_),next;m/test2/ and $y=1,push(@a,$_),next;END{if ($x&&$y) {print for (@a)}}' myfile

Regards!

...JRF...

Re: egrep and patterns

ghostdog74 — Fri, 05 Jun 2009 23:43:40 GMT

you just use awk

awk '/test1/{s[FNR]=$0}/test2/{f=1;s[FNR]=$0}END{if (f){for(i in s) print s[i]}}' file

Re: egrep and patterns

Dennis Handly — Sat, 06 Jun 2009 10:27:18 GMT

>patterns are not in the same line

You can use multiple passes of grep like JRF said:
grep test1 $(grep -l test2 file-list)

Re: egrep and patterns

emha_1 — Sat, 06 Jun 2009 15:07:22 GMT

Hi,

what about simple:

grep -e test1 -e test2

emha.

Re: egrep and patterns

ghostdog74 — Sat, 06 Jun 2009 23:28:58 GMT

>grep -e test1 -e test2

if file only has test1 and not test2, then it will not work.

Re: egrep and patterns

emha_1 — Tue, 09 Jun 2009 11:02:32 GMT

>if file only has test1 and not test2, then it will not work

you are right, I'm wrong.
as the grep is line matching tool, I haven't realized original question was 'to find the FILE that contains both the lines test1 and test2'.
I understood it to list lines from input that contains test1 and test2.

emha.