shell script -parsing a log file

jd · 10-01-2002 2:35am #1

Any one some quick ideas on this?
I basically want to find the first occurrence of a pattern in a log file and then manipulate the rest of the file from there on eg a date stamp-dumping into another temp file would be ok..
jd

ecksor · 10-01-2002 2:58am

I'm almost certain that sed can do this by specifying the pattern you want as the first range parameter to whatever operator you're using. That may or may not be helpful though.

something like

cat $filename |
( flag=""
  while read logline
  do
       if echo $logline | grep "pattern" > /dev/null
       then
             flag="X"
       fi

        if [ "$flag" = "X" ]
        then
             echo $logline
        fi
   done) | tr a A

Would turn all lowercase a to A in a file starting with the line where "pattern" occured.

Wookie · 10-01-2002 11:37am

Given a file called test.txt

wookie
wookie
marker
hello world
hello world

run test.com

T1=`grep -n marker ./test.txt | cut -f 1 -d : -`
T2=`cat test.txt | wc -l`
T3=$(($T2 - $T1))
tail -$T3 test.txt > new.txt

new.txt should then contain the following

hello world
hello world

I think this does what you need. However it is probably two much of a hack to ever be used.

jd · 10-01-2002 10:33pm

Originally posted by Wookie
Given a file called test.txt

wookie
wookie
marker
hello world
hello world

run test.com

T1=`grep -n marker ./test.txt | cut -f 1 -d : -`
T2=`cat test.txt | wc -l`
T3=$(($T2 - $T1))
tail -$T3 test.txt > new.txt

new.txt should then contain the following

hello world
hello world

I think this does what you need. However it is probably two much of a hack to ever be used.

yes that was something like I was trying..
I made one change to make it work though

T1=`grep -n marker ./test.txt | cut -f 1 -d : - |head -1 - `
tx
jd

shell script -parsing a log file

Comments