CODEWITHSUNDEEP
linuxshell
mayank arora
mayank arora

Reputation: 117

Repetitive sed use on same file

Below is the shell script I am using to replace 200 plus strings in nagios.cfg.

I want sed to replace all 200 strings in same file but instead its copying the content 200 times with each having one change.

    #set -x
    cat partthree | while read  LINE
    do hostnam=`echo $LINE | awk '{print $1}'`
    echo $hostnam
    a=`less hosts | grep $hostnam -A5 | grep -i md5sum | awk '{print $2}'`
    b=`less partthree  | grep $hostnam | awk '{print $2}'`
    echo $a
    echo $b
    sed  "s/$a/$b/g" hosts >> sample
#   sed  "s/$a/$b/g" hosts > sample will change only last occurence
    done

Tried using sed -i "s/$a/$b/g" hosts directly

@Expected output:

define host {
        address        sampleaddress
        alias           samplealaias
        host_name       samplealhostname
        _md5sum         def(with replaced string)

}

Output

define host {
   address         sampleaddress
   alias           samplealaias
   host_name       samplealaias
   c8h(origanal string)         def(string that is supposed to replace original c8h value)

}

Upvotes: 1

Views: 152

Answers (1)

Flopp
Flopp

Reputation: 1947

The following script should do the job. It starts with the initial hosts file and performs the individual replacements one after another. The main difference to the original script is that the replacement results are not appended to the result file, but are applied in-place (sed -i) on the result file.

# start with 'sample' = 'hosts'
cp hosts sample

# perform replacements one after another in 'sample'
cat partthree | while read LINE ; do
    hostnam=`echo $LINE | awk '{print $1}'`
    a=`less hosts | grep $hostnam -A5 | grep -i md5sum | awk '{print $2}'`
    b=`less partthree  | grep $hostnam | awk '{print $2}'`
    # modify 'sample' in-place
    sed -i "s/$a/$b/g" sample
done

Upvotes: 1

Related Questions