Filter block of output from a log file

gauravjain21 · #1 (**permalink**) 06-25-07, 16:34

Hi,

I've a log file with lot of data in it. like

Hello how is everyone
Hope all is well
-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
yahooooo
gooooooogle 56
ning 8
utube
technorati
-*-*-*-*-*- RULE EXECUTION -*-*-*-*-*-
life couldn't be better
thank you for your response and help

I need data or block of content appearing as is between the lines "-*-*-*-*-*- OUTPUTS -*-*-*-*-*-" and "-*-*-*-*-*- RULE EXECUTION -*-*-*-*-*-"

radoulov · #2 (**permalink**) 06-25-07, 17:29

Code:

awk '$0~/OUTPUTS/,$0~/RULE EXECUTION/{if($0~/OUTPUTS|RULE EXECUTION/)next;else print}' infile

gauravjain21 · #3 (**permalink**) 06-25-07, 22:56

Log file is like this - I am interested in data between first occurrence (from top) of OUTPUT and RULE EXECUTION.

Answer should be :

yahooooo
gooooooogle 56
ning 8
utube
technorati

Log File

-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello
-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello
-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello
-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello
-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello

-*-*-*-*-*- OUTPUTS -*-*-*-*-*-

yahooooo
gooooooogle 56
ning 8
utube
technorati

-*-*-*-*-*- RULE EXECUTION -*-*-*-*-*-

-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
counts
-*-*-*-*-*- RULE EXECUTION -*-*-*-*-*-

-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello
-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
hello

-*-*-*-*-*- OUTPUTS -*-*-*-*-*-
counts
-*-*-*-*-*- RULE EXECUTION -*-*-*-*-*-

radoulov · #4 (**permalink**) 06-26-07, 06:46

Code:

awk '{gsub(/.*\*-\n|\n-\*.*/,"",$1);print $1}'  FS="RULE EXECUTION" RS="~" infile

Use nawk on Solaris.

gauravjain21 · #5 (**permalink**) 06-26-07, 14:34

awk '{gsub(/.*\*-\n|\n-\*.*/,"",$1);print $1}' FS="RULE EXECUTION" RS="~" infile

awk: Input line -*-*-*-*-*-*-*-*-*-* cannot be longer than 3,000 bytes.
The source line number is 1.

Tyveleyn · #6 (**permalink**) 06-26-07, 15:04

What about:

Code:

awk '$0 ~ /RULE EXECUTION/  {f = 0}              
    f == 1                  {print $0}           
    $0 ~ /OUTPUTS/          {f = 1}' infile

Regards

gauravjain21 · #7 (**permalink**) 06-26-07, 16:23

That gives the full log file. I am interested in the data between first occurrence (from top) of OUTPUT and RULE EXECUTION. Thats it, nothing else.

Tyveleyn · #8 (**permalink**) 06-26-07, 18:17

Must be the awk implementation then. It works perfectly well on my system, using mawk. But let's do something else then; if you only need the lines between the first occurrences of 'OUTPUTS' and the subsequent 'RULE EXECUTION' it's probably faster to use:

Code:

sed -e '1,/OUTPUTS/d' -e '/RULE EXECUTION/,$d' infile

The semantics of these sed expressions are universal.

Regards

radoulov · #9 (**permalink**) 06-27-07, 01:27

Quote:

Originally Posted by gauravjain21

awk '{gsub(/.*\*-\n|\n-\*.*/,"",$1);print $1}' FS="RULE EXECUTION" RS="~" infile

awk: Input line -*-*-*-*-*-*-*-*-*-* cannot be longer than 3,000 bytes.
The source line number is 1.

I see ...
Try this:

Code:

awk '{rec[NR]=$0}
/OUTPUTS/{start=NR}
/RULE EXECUTION/{end=NR;exit}
END{for(i=start+1;i<end;i++)print rec[i]}' infile

gauravjain21 · #10 (**permalink**) 06-27-07, 22:43

This script rocks - Thanks much.

Internet Services

Web Development

Digital Marketing

Consumer Tech