Question

I have the following text file for which I need to compare values from each line, namely items 2-4 against items 5-7. I’m stuck with bash/awk/sed on this one.

Sample data:

[hartford tmp]$ cat flist
a1 1 2 3 x y z
b1 3 2 1 z y x
c1 1 2 3 1 2 3
d1 4 5 6 6 5 4
e1 a b c a b c
f1 x y z x y z

It works with the following script but its just unbearably slow, probably because all of the echos.

[hartford tmp]$ cat pdelta.sh
#!/bin/bash

cat flist |while read rec; do
    f1="$(echo $rec | awk '{ print $1 }')"
    f2="$(echo $rec | awk '{ print $2 }')"
    f3="$(echo $rec | awk '{ print $3 }')"
    f4="$(echo $rec | awk '{ print $4 }')"
    f5="$(echo $rec | awk '{ print $5 }')"
    f6="$(echo $rec | awk '{ print $6 }')"
    f7="$(echo $rec | awk '{ print $7 }')"

    if [[ "x${f2} x${f3} x${f4}" != "x${f5} x${f6} x${f7}" ]]; then
            echo "$f1 DOES NOT MATCH"
    fi
done

When run, the output is exactly what I’m looking for but it’s too slow when dealing with a file that’s 50k+ lines long.

[hartford]$ ./pdelta.sh
a1 DOES NOT MATCH
b1 DOES NOT MATCH
d1 DOES NOT MATCH

What is a more efficient way to accomplish this?

Answer 1

You can use awk to output all the matching id’s:

awk '{ if ($2 == $5 && $3 == $6 && $4 == $7) { print $1 } }' < flist