I have two files:

$ cat inds
2
5
3

$ cat family
1 2 3 4
1 5 3 4
1 3 0 0
1 4 0 0
1 7 0 0
1 6 2 7
1 8 2 7

I want to write a new file with all the records in family that are NOT in inds, but matching only in field 2 of family. I foolishly tried

grep -v -f inds family

But of course it matches anywhere in the line, and it removes the last two lines as well because they have the 2 in field 3.

awk '{if($2!="2" && $2!="3" && $2!="5") print $0}' family

would work if my real inds file was not hundreds and hundreds of lines long... So basically, I am looking for something that can inverse-match every pattern from a long long list of patterns, but only for field 2.

I hope this makes sense. Thank you.

Among many ways to accomplish that, here's a quickie that will work in awk:It reads the entire inds file into an associative array using the first field as the index and setting an arbitrary value, then for each line in the second file (all subsequent files, actually) it checks to see whether field two exists as an index in the inds array.

1 members found this post helpful.

Thank you, this is exactly what I was looking for!

I think I even understand how it does it.

I am assuming rk is presenting a full layman version in case you are new, but here is the same code with the superfluous parts removed:

1 members found this post helpful.

The assumption was correct, I am new. Thank you for the short version, with some googling I think I understand what is going on.