CODEWITHSUNDEEP
bashsedawkscriptinggrep
imre
imre

Reputation: 489

Filtering throung a listed text

I have a list of text that looks like this:

function.o U out
function.o T fail
function.o U main
myPart.o T out
myPart.o U fail

and so on...

I want the final list to look like this:

The .o file with U beside it : The .o file with T beside it /the third column they have both/

So the example up there would look like this filtered:

function.o : myPart.o /out/
myPart.o : function.o /fail/

So I list only the lines that have the same symbol in the third column (like out, fail and so on). Until now I came up with this:

awk '$3 in t{ln[$3]=t[$3]" -> "$1;next}{t[$3]=$1}END{for(s in ln) print ln[s],"("s")"}' "$file"

But that prints everything without accounting for the U/T symbol. Any ideas please?

Upvotes: 1

Views: 48

Answers (2)

jaypal singh
jaypal singh

Reputation: 77105

$ awk '
$3 in t{line[$3]=($2=="U")?$1" : "t[$3]:t[$3]" : "$1;next}{t[$3]=$1}
END{for(k in line) print line[k],"/"k"/"}
' file
myPart.o : function.o /fail/
function.o : myPart.o /out/

Upvotes: 1

Etan Reisner
Etan Reisner

Reputation: 80931

Does this do what you want?

{
    file[$3,$2] = $1
    thirds[$3]++
}

END {
    for (third in thirds) {
        if (file[third,"T"]) {
            print file[third,"T"], ":", file[third,"U"], "/" third "/"
        }
    }
}

Upvotes: 0

Related Questions