I have a task: Write a script for the summation of integers stored in the file. Form a call script: example: sum a.txt 3 4

The input file can contain several columns of integer. The individual columns are separeted by speces or tabs. The script should sum a appropriate columns and write the result to stdout. So when we have sum a.txt 3 4 we need to add the number of the third and fourth columns file.

So I do this:

#!/bin/bash
array1=( "$@" )
let LA=${#array1[@]}-1
awk '{for(i=1;i<=$LA;i++)y+=$'${array1[i]}'; print y}' a.txt

but I have an error: awk: : 1unexpected character '.'

Please help is there another way to add up the number of columns whose number are given in the procedure call script?

On this forum a got answer to do this the following:

#!/bin/bash
awk -v col1=$2 -v col2=$3 '{sum1 += $col1; sum2 += $col2} END{print sum1,sum2}' $1

but what if we don't know the amount of numbers of columns that will be given in the procedure example: ./sum a.txt 2 3 ... n (maybe I need use for but how?)

You do not need AWK for this. Bash is enough:

$ cat data.txt
1 2 3 4 5 6
7 8 9 10 11 12
13 14 15 16 17 18
$ ./csum.sh $(seq 0 5) < data.txt
21 24 27 30 33 36
$ cat csum.sh
#! /bin/bash
SUM=()
S=0
for N in "$@"; do
  SUM[$S]=0
  ((S++))
done
while IFS=$'\n' read LINE ; do
  COLS=($LINE)
  S=0
  for C in "$@"; do
    SUM[$S]=$(expr ${SUM[S]} + ${COLS[C]})
    ((S++))
  done
done
echo ${SUM[*]}