CODEWITHSUNDEEP
shellsortingcountawkuniq
saggi
saggi

Reputation: 1

How do I count grep results for a string, but specify exclusions?

I have maillog file with below parameters

relay=mx3.xyz.com

relay=mx3.xyz.com

relay=mx1.xyz.com

relay=mx1.xyz.com

relay=mx2.xyz.com

relay=home.xyz.abc.com

relay=127.0.0.1

I want to count all relay except 127.0.0.1

Output should like this

total relay= 6

mx3.xyz.com = 2

mx1.xyz.com = 2

mx2.xyz.com = 1

home.xyz.abc.com = 1

Upvotes: 0

Views: 134

Answers (3)

Faiz
Faiz

Reputation: 16255

If you don't mind using awk:

awk -F= '$2 != "127.0.0.1"  &&  /relay/ {count[$2]++; total++} 
  END { print "total relay = "total; 
        for (k in count) { print k" = " count[k]}
      }' maillog

And you could also make do with just uniq and grep, though you won't get your total this way:

 grep relay maillog  | cut -d= -f2 | grep -v 127.0.0.1 | uniq -c

And if you don't hate perl:

perl -ne '/relay=(.*)/ and $1 ne "127.0.0.1" and ++$t and $h{$1}++;
  END {print "total = $t\n"; 
       print "$_ = $h{$_}\n" foreach keys %h;
  }' maillog

Upvotes: 3

glenn jackman
glenn jackman

Reputation: 246807

I would definitely use awk for this (@Faiz's answer). However I worked out this excruciating pipeline 

cut -d= -f2 filename | grep -v -e '^[[:space:]]*$' -e 127.0.0.1 | sort | uniq -c | tee >(echo "$(bc <<< $(sed -e 's#[[:alpha:]].\+$#+#' -e '$a0')) total") | sed 's/^ *\([0-9]\+\) \(.*\)/\2 = \1/' | tac

outputs

total = 6
mx3.xyz.com = 2
mx2.xyz.com = 1
mx1.xyz.com = 2
home.xyz.abc.com = 1

Please do not upvote this answer ;)

Upvotes: 0

Kent
Kent

Reputation: 195049

here you go:

 awk -F= '$2!="127.0.0.1"&&$2{t++;a[$2]++} END{print "total relay="t; for(x in a)print x"="a[x]}' yourfile

the output would be:

total relay=6
mx2.xyz.com=1
mx1.xyz.com=2
mx3.xyz.com=2
home.xyz.abc.com=1

Upvotes: 2

Related Questions