Hi,
I want to compare two columns and find out missing entries e:g
Column 1 Column 2
1 1
2 2
3 13
4
10
19
234
Results woud be 13. I will appreciate very much if anyone help me :).
awk '$NF != $(NF-1){print $NF}' file
perl -lane 'print $F[1] if ( $F[1] && $F[1] != $F[0])'
Or:
awk '$0=$1==$2?0:$2' infile
Else Or:
perl -ane '$F[1]!=$F[0]&&print$F[1]' file
Or:
perl -lane'print$F[1]if$F[1]&&!/(\S+)\s+\1$/' infile
Edit: Re-reading the initial question, I realized that the OP didn't specify how multiple values should be displayed 
So I'm just guessing ...
I pretty much meant what I posted, but this will add a new line after every match, which is probably a good thing:
perl -ane '$F[1]!=$F[0]&&$F[1]&&print$F[1],"\n"' dat
Yes, and you can use the -l option for that, like you did i your first post to this thread.
I mean this:
% perl -lane '$F[1]!=$F[0]&&print$F[1]' infile
13
%
vs. this:
% perl -lane'print$F[1]if@F[1]&&!/(\S+)\s+\1$/' infile
13
%
Sorry, if it was not clear.
You are mixing and matching my posts.
The second post, the one you are referring to, omits the -l flag, which prevents the problem with the extra output lines, as you show there. you include the -l flag, hence the problem.
Try it exactly the way I posted it, with -ane
perl -ane '$F[1]!=$F[0]&&print$F[1]' file
Pretty sure you do get not all the empty lines, that way.... let me know.
I think the third attempt is best anyhow.
Thanks, and happy Perlin'
Scott
Hi Scott,
You're right, I was not clear at all.
Correct, you get the value (and eventually the values concatenated together with no separator) without the trailing newline character.
Yes,
with the original sample I get this:
% perl -ane '$F[1]!=$F[0]&&print$F[1]' infile
13%
With the test input file I was using (more than one record with mismatching fields) I got:
% cat infile
1 1
2 2
3 13
4
10
19 1
234
% perl -ane '$F[1]!=$F[0]&&print$F[1]' infile
131%
Correct, sorry for the noise ...
Better than this?
perl -lane '$F[1]!=$F[0]&&$F[1]&&print$F[1]' infile
Regards
Dimitre
When I say "my third attempt" I mean the one-liner shown here, it accounts for duplicates, extra white space and outputs the line feed after every match.
[root@localhost cols]# cat dat
1 1
2 2
3 13
4
10
19
234
3 14
4 4
5 5
112 13
[root@localhost cols]# perl -ane '$F[1]!=$F[0]&&$F[1]&&print$F[1],"\n"' dat
13
14
13
I guess when you add the -l back in, the "\n" becomes unnecessary, so: no, not better than "that". You are correct
Thanks for your reply. My sample input was not clear. Duplicate values can occur anywhere in the 2nd column than side by side. e:g
10 1
1 4
2 13
3
4
93
So how the output should look like?