Add Columns Values with Shell - shell

Question

I've an input file which looks as below.
pmx . pmnosysrelspeechneighbr -m 1 -r
INFO: The ROP files contain suspected faulty counter values.
They have been discarded but can be kept with pmr/pmx option "k" (pmrk/pmxk) or highlighted with pmx option "s" (pmxs)
Date: 2017-11-04
Object Counter 14:45 15:00 15:15 15:30
UtranCell=UE1069XA0 pmNoSysRelSpeechNeighbr 0 1 0 0
UtranCell=UE1069XA1 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XA2 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XA3 pmNoSysRelSpeechNeighbr 0 0 2 0
UtranCell=UE1069XB0 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XB1 pmNoSysRelSpeechNeighbr 0 0 0 3
UtranCell=UE1069XB2 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XB3 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XC0 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XC1 pmNoSysRelSpeechNeighbr 0 0 0 4
UtranCell=UE1069XC2 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1069XC3 pmNoSysRelSpeechNeighbr 0 0 1 0
UtranCell=UE1164XA0 pmNoSysRelSpeechNeighbr 0 3 0 0
UtranCell=UE1164XA1 pmNoSysRelSpeechNeighbr 0 0 0 0
UtranCell=UE1164XA2 pmNoSysRelSpeechNeighbr 1 0 0 0
Now I want the output as below which is basically sum of the time column (from $3 to $6) values.
Counter 14:45 15:00 15:15 15:30
pmNoSysRelSpeechNeighbr 1 4 3 7
I've been trying with below command. But it's just giving sum of one column values:
pmx . pmnosysrelspeechneighbr -m 1 -r | grep - i ^Object| awk '{sum += $4} END {print $1 , sum}'

Answer 1

Try this out, you will get both header and trailer as sum of individual columns.
BEGIN {
trail="pmNoSysRelSpeechNeighbr";
}
{
if($1=="Object") print $2 OFS $3 OFS $4 OFS $5 OFS $6;
else if($1 ~ /^UtranCell/) {
w+=$3; x+=$4; y+=$5; z+=$6;
}
}
END {
print trail OFS w OFS x OFS y OFS z;
}

Answer 2

You need to sum each of the columns separately:
awk -v g=pmNoSysRelSpeechNeighbr '$0 ~ g { for(i=3;i<=6;i++) sum[i]+=$i }
END { printf g; for(i=3;i<=6;i++) printf OFS sum[i] }' file
but only for lines (records) containing the group (counter) of interest ($0~"pmNoSysRelSpeechNeighbr").
Note you (almost) never need to pipe grep's output to awk, because awk already supports extended regular expressions filtering with /regex/ { action }, or var ~ /regex/ { action }. One exception would be the need for PCRE (grep -P).
As an alternative to awk for simple "command-line statistical operations" on textual files, you could also use GNU datamash.
For example, to sum columns 3 to 6, but group by column 2:
grep 'UtranCell' file | datamash -W -g2 sum 3-6