Find differences exist in A but not B

This is A:

SomeFollowingText 12-45-78-54 1.1.1.1
SomeFollowingText 78-56-14-54 2.2.2.2
SomeFollowingText 52-12-98-ab 3.3.3.3
SomeFollowingText 89-za-gg-99 4.4.4.4

This is B:

78-56-14-54
ab-g2-ax-47
52-12-98-ab
g2-47-n7-o9
56-13-z8-ab

The expected output:

SomeFollowingText 12-45-78-54 1.1.1.1
SomeFollowingText 89-za-gg-99 4.4.4.4

This is my failed attempt:

awk '{print $2}' /home/path/A | while read line
do
    if [ "grep $line /home/path/B" != "$line" ]
    then
        echo $line
    fi
done >> /home/path/non-match

But the non-match contains the whole A.

I also tried this:

cat /home/path/A | while read line
do
    if [ "$(grep $line /home/path/B)" != "$line | awk '{print $2}'" ]
    then
        echo $line
    fi
done >> /home/path/non-match

$ awk 'FNR==NR{a[$1];next} !($2 in a)' B A
SomeFollowingText 12-45-78-54 1.1.1.1
SomeFollowingText 89-za-gg-99 4.4.4.4

if [ "grep $line /home/path/B" != "$line" ] compares two strings, one is e.g. grep 12-45-78-54 /home/path/B and the other is 12-45-78-54. Those will always be different.

What you apparently want to do is compare the output of the grep command to $line, so you should use if [ "$(grep $line /home/path/B)" != "$line" ] instead.

However, this will not give you the expected output, since you effectively deleted the 1st and 3rd token from each input line using awk and you can't recover those from within the loop.

A much simpler approach to your problem would be grep -vf B A >non-match, as @tshiono suggested.

You can do this pretty easily in awk:

awk '(NR==FNR) {b[$1]=$1}; (NR>FNR && ! b[$2]) {print}' /home/path/B /home/path/A >/home/path/non-match

Explanation: awk reads through all lines in all supplied input files (/home/path/B and A, in that order). FNR is the line ("record" number) in the current file, and NR is the overall line number, so comparing them is an easy way to tell if it's processing the first file (they'll be equal) or not (unequal).

(NR==FNR) {b[$1]=$1} means: for each line in the first file (/home/path/B) add the first (only) field to an array named b, indexed by itself.

(NR>FNR && ! b[$2]) {print} means: for each line in the second file (/home/path/A), if the second field ($2) isn't in the b array, print the line.