Extract lines between two patterns from a file

Question

I need to extract particular lines between two matching patterns from a huge file.

Let's say pattern1 (unique in a file) matches a particular line # n and pattern2 (not unique in a file) matches line # m next immediate match after line # n. Then I want to extract all lines between and including line #n to #m

Sample File contents

***************************************************************************
text line # n-2
text line # n-1
********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************
text line # m+1
text line # m+2
***************************************************************************

Desired Output

********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************

It would be appreciated if you could help me solve this problem

I marked this question as a duplicate to the one I opened to cover all cases. This way, we have a main place with good explanations on this topic. — fedorqui, Mar 27 '17 at 13:28

fedorqui · Accepted Answer · 2013-10-04T10:01:22.050

24

This can be an approach:

$ awk '/pattern1/ {p=1}; p; /pattern2/ {p=0}' file
********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************

When it finds pattern1, then makes variable p=1.
it just prints lines when p==1. This is accomplished with the p condition. If it is true, it performs the default awk action, that is, print $0. Otherwise, it does not.
When it finds pattern2, then makes variable p=0. As this condition is checked after p condition, it will print the line in which pattern2 appears for the first time.

If you want an exact match of the lines:

$ awk '$0=="pattern1" {p=1}; p; $0=="pattern2" {p=0}' file

Test

$ cat a
***************************************************************************
text line # n-2
pattern1
********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************
pattern2
text line # m+2
pattern2
***************************************************************************
$ awk '/pattern1/ {p=1}; p; /pattern2/ {p=0}' a
pattern1
********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************
pattern2

edited Oct 04 '13 at 10:01

answered Oct 04 '13 at 09:29

fedorqui

275,237
103
548
598

Thanks for your response. but I can't match line # n-1 and line # m+1. They are not deterministic – jkshah Oct 04 '13 at 09:39
@jkshah oh, now I see. Check my updated answer. I take into account the fact that you are looking for a pattern1, see the test. – fedorqui Oct 04 '13 at 09:44
updated ans excludes line #n and #m, which is not desired. Can you please check? It seems your solution in line with what is suggested by @Jotne and it works fine. – jkshah Oct 04 '13 at 09:50
It does not exclude line #n and #m. This solution provides a more general approach for you looking for specific pattern1 and pattern2. – fedorqui Oct 04 '13 at 09:53
I substituted `Results` for `pattern1` and `^[*]+$` for `pattern2` and it removed starting and ending lines. Am I doing something worg? bdw +1 for your explanations – jkshah Oct 04 '13 at 09:56
Ok, I was getting confused with the fact of printing `n` or `n+1`. Now I turned the condition so both lines are printed: the one with `pattern1` and the first with `pattern2`. Hope it is clear now :) – fedorqui Oct 04 '13 at 10:02
1

Now that's perfect and concise. Thanks a lot! – jkshah Oct 04 '13 at 10:11

score 8 · Answer 2 · edited Jan 08 '17 at 11:51

8

Using sed:

$ sed '/start_pattern_here/,/end_pattern_here/!d' inputfile

and in OP's specific case:

$ sed '/[*]* Results [*]*/,/^[*]*$/!d' inputfile
********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************

Assuming that the unique pattern was *** Results *** and the non-unique one was ********.

edited Jan 08 '17 at 11:51

einpoklum

118,144
57
340
684

answered Oct 04 '13 at 09:38

devnull

118,548
33
236
227

Thanks for your response. This works pretty fine. – jkshah Oct 04 '13 at 09:43
@devnull, do you believe this solution preferable to the awk soluton? If so, can you explain why? – einpoklum Jan 08 '17 at 11:49

score 1 · Answer 3 · answered Oct 04 '13 at 09:32

1

Using awk

awk '/Result/ {p=1;print;next}  /^\*\*\*\*\*/ && p {p=0;print} p' file
********************************* Results *********************************
SUCCEEDED
...
...
some text
***************************************************************************

answered Oct 04 '13 at 09:32

Jotne

40,548
12
51
55

Thanks for your response. This works pretty fine. – jkshah Oct 04 '13 at 09:45

Extract lines between two patterns from a file

3 Answers3

Test

Linked

Related