mjhjmtu

I have the following input:

1. Track 01 05:21
2. Track 02 07:39 
3. Track 03 04:27

I came up with the following script

awk -F: 'a+=$(NF-1)*60+$(NF);printf"%d:%02d:%02dn",a/3600,a%3600/60,a%3600%60' file

But the result I get is this:

0:01:21
0:04:00
0:07:27

In turn, with the same script, when I have a file with only the time and nothing else I get the correct output

Input:

05:21
07:39
04:27

Output:

0:05:21
0:13:00
0:17:27

Use two awk commands

awk 'print $NF' file |awk 'a+=$(NF-1)*60+$(NF);printf"%d:%02d:%02dn",a/3600,a%3600/60,a%3600%60'

the first awk extract the las field of the file and the second awk makes the procees you want.

this works whether the file contets only the time or not, as long as the time is the last field in the line

This is the problem you're having: when FS=":", your first field is "1. Track 01 05", and when you use that string in an arithmetic context, the first non-digit truncates the number, so the minute value is "1". You need to use space or colon as the field separator, but then any trailing whitespace means you have extra empty fields.

I would do

awk '
 BEGIN h = m = s = 0
 function add_time(min, sec) 
 s += sec
 m += min
 while (s >= 60) s -= 60; m++
 while (m >= 60) m -= 60; h++
 
 
 split($NF, t, /:/)
 add_time(t[1], t[2])
 printf "%d:%02d:%02dn", h, m, s
 
' file