How to capture 1000 lines before a string match and 1000 line a string match including line of string match ?

sysmicuser · 11-11-2017, 06:12 AM

Hey Guys,

Not sure if this is possible but I have a requirement, imagine a log file with content below

...1000 + lines
aaaknfdnfnjwefjn jwhbfewhjfb
cbcjhbdkqwhjwfl jhfwhfwjhqwjrie
TVSFSSJ.csv
whugefiuwqehhfowijf
khcbkjsdbkqjwfwhfej uihfowefoiwehjf
jwfvjwebfiwefwef nwehfiweuhfiewuhfi
....1000 + lines

What I am interested is to extract 1000 lines before "TVSFSSJ.csv" pattern match line, line with "TVSFSSJ.csv" and 1000 lines after "TVSFSSJ.csv" pattern match line.

Is this achievable ?

Please assist. Thanks

TenTenths · 11-11-2017, 06:24 AM

Code:

man grep

Hint, look for "Context Line Control"

syg00 · 11-11-2017, 06:24 AM

Of course it's achievable. As a search would tell you.
"man grep".

sysmicuser · 11-11-2017, 06:45 AM

Thank you. I did man grep and answer is -A and -B but I still have two more questions. I want the line from the first time pattern match and 1000 lines before and after the first pattern match. Secondly, I believe -A and -B does not give me the pattern match line. How do we get that ?

TenTenths · 11-11-2017, 06:48 AM

Quote:

Originally Posted by sysmicuser

Secondly, I believe -A and -B does not give me the pattern match line.

Did you actually bother to try it? Your belief is wrong.

sysmicuser · 11-12-2017, 02:19 AM

Quote:

Originally Posted by TenTenths

Did you actually bother to try it? Your belief is wrong.

Yep, it gives me pattern match line BUT my requirement is yet not fulfilled. Whenever there is a pattern match I want grep to be 1000 lines before and after for only first occurrence when you read the file top to down. Unfortunately it is giving for all occurrences which doesn't help me. I tried with

Code:

-m 1

switch however then line AFTER match are not displayed..

Turbocapitalist · 11-12-2017, 02:49 AM

The -m option does not interfere with either -A or -B.

So please post the actual options you are using as well as the version: grep -V.

sysmicuser · 11-12-2017, 04:33 AM

Quote:

Originally Posted by Turbocapitalist

The -m option does not interfere with either -A or -B.

So please post the actual options you are using as well as the version: grep -V.

Answer to first question.
Just grep without -m option

Code:

[root@centos6 bin]# grep -A 2 -B 2 "SampleTextFile.txt" /var/log/messages
Nov 10 00:14:54 centos6 kernel: Kernel logging (proc) stopped.
Nov 10 00:14:54 centos6 rsyslogd: [origin software="rsyslogd" swVersion="5.8.10" x-pid="1513" x-info="http://www.rsyslog.com"] exiting on signal 15.
File SampleTextFile.txt
First Line after original SampleTextFile.txt
Nov 12 18:13:48 centos6 kernel: imklog 5.8.10, log source = /proc/kmsg started.
Nov 12 18:13:48 centos6 rsyslogd: [origin software="rsyslogd" swVersion="5.8.10" x-pid="1519" x-info="http://www.rsyslog.com"] start
--
Nov 12 18:14:19 centos6 polkitd[2382]: started daemon version 0.96 using authority implementation `local' version `0.96'
Nov 12 18:14:19 centos6 rtkit-daemon[2388]: Sucessfully made thread 2386 of process 2386 (/usr/bin/pulseaudio) owned by '42' high priority at nice level -11.
File SampleTextFile.txt
Second Line after original SampleTextFile.txt
Nov 12 18:14:19 centos6 pulseaudio[2386]: alsa-util.c: Disabling timer-based scheduling because running inside a VM.
Nov 12 18:14:19 centos6 rtkit-daemon[2388]: Sucessfully made thread 2392 of process 2386 (/usr/bin/pulseaudio) owned by '42' RT at priority 5.
[root@centos6 bin]#

Now when I use -m option

Code:

[root@centos6 bin]# grep -m 1 -A 2 -B 2 "SampleTextFile.txt" /var/log/messages
Nov 10 00:14:54 centos6 kernel: Kernel logging (proc) stopped.
Nov 10 00:14:54 centos6 rsyslogd: [origin software="rsyslogd" swVersion="5.8.10" x-pid="1513" x-info="http://www.rsyslog.com"] exiting on signal 15.
File SampleTextFile.txt
[root@centos6 bin]#

Lastly grep version 2.20

Code:

[root@centos6 bin]# grep -V
grep (GNU grep) 2.20
Copyright (C) 2014 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>.
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.

Written by Mike Haertel and others, see <http://git.sv.gnu.org/cgit/grep.git/tree/AUTHORS>.
[root@centos6 bin]#

Thanks in advance

sysmicuser · 11-13-2017, 04:44 PM

Any update from any learned guru ?

AwesomeMachine · 11-13-2017, 05:50 PM

You're not doing it right.

Code:

# grep -m 1 -A 2 -B 2 "SampleTextFile.txt" /var/log/messages

is looking for the string SampleTextFile.txt in /var/log/messages. You want

Code:

$ grep -m 1 -A 1000 -B 1000 some_string some_file

It works fine if you don't specify a file in place of the string, unless you want to find the file name in the specified file. I direct you back to

Code:

$ man grep

or a grep tutorial online, because you don't have a grasp on the basic usage.

sysmicuser · 11-14-2017, 01:55 AM

I'm doing it right and I am indeed grepping for a file name in /var/log/messages. Suggest you try exactly what I am doing and advise me if I am wrong.

I fail to understand what am I doing wrong..

TenTenths · 11-14-2017, 05:08 AM

Quote:

Originally Posted by Turbocapitalist

The -m option does not interfere with either -A or -B.

Looks like it does when the matching pattern appears in the -B block.

For example:

Code:

$ cat test
1
2
3
A
4
B
4
C
4
5
6
7
8
9
10

What the OP is expecting:

Code:

$grep -m 1 -A 2 -B 2 4 test Search for the first occurrence of 4
3
A
4
B
4

What actually happens in grep 2.20 is what the OP is observing:

Code:

$ grep -m1 -A 2 -B 2 4 test
3
A
4
B

So it does appear that the -m option does indeed interfere with -A

Turbocapitalist · 11-14-2017, 05:21 AM

Quote:

Originally Posted by TenTenths

Looks like it does when the matching pattern appears in the -B block.

Ah. I had tested several permutations but not that one. I suspect there may be no work-around with grep. There may be a short way to do it in AWK or perl.