Matrix addition in awk

Question

I have a bunch of variables which look like this:

(DURATION 1.57) + (DURATION 2.07)
(T0 10) (T1 0) (TX 0) + (T0 12) (T1 0) (TX 1)
(TC 1) (IG 0) + (TC 2) (IG 3)

Is it possible to have awk process this such that the result is:

(DURATION 3.64)
(T0 22) (T1 0) (TX 1) 
(TC 3) (IG 3) 

Or can anyone recommend another unix program I can use to do this?

Answer 1

Here is one way to do it:

awk '{
    gsub(/[()+]/, "")
    for(nf=1; nf<=NF; nf+=2) {
        flds[$nf] += $(nf+1)
    }
    sep = ""
    for(fld in flds) {
        printf "%s(%s %g)", sep, fld, flds[fld]
    sep = FS
    }
    print "";
    delete flds
}' file
(DURATION 3.64)
(T0 22) (T1 0) (TX 1)
(TC 3) (IG 3)

• We remove the special characters ()+ using gsub() function.
• We iterate over all fields adding variables to an array and adding the values
• We iterate over the array, printing them in our desired format.
• Add a new line after we are done printing
• Delete the array so that we can re-use it on next line
• Note: The order of lines will be same as input file but using the in operator for our for loop the variables on each line may appear in random order.