How to separate a variable into multiple variable in bash

Question

#! /bin/bash
readarray rows < Process.dat

for ((i=0, j=1; i< ${#rows[@]}; i++, j++)); do
   declare "row$j"="${rows[$i]}"

done

echo "$row1"

the output for this code is 10110 556 4433, I have tried using the same piece of code again changing the variables of "rows" and "row" however there is no ouput when i do this, How can i split up the variable "row1" into the 3 seperate numbers each assigned to a different variable, e.g row1_1 = 10110, row1_2 = 556, row1_3 = 4433

the input file consists of 5 rows and 3 columns of numbers:

10110 556 4433
10120 424 54
12452 123 534
22042 432 12
25321 423 73

The columns represent index time, burst time and id number respectively.

I need to work out:

• execution order,
• average wait time,
• average completion time

e.g execution order 4433>54>534>12>73. 
average completion time = 6132/5
(556+(556+424)+(556+434+123)+(556+424+123+432)+(556+424+123+432+423)) / 5

average wait time = 4174/5
(0+556+(556+424)+(556+424+123)+(556+424+123+432)) / 5

the endgame is to have each of these numbers assigned to a unique variable that I can later use to do math operations

output from od -c Process.dat

0000000   0   1   0   1   3   4       7   5   6       4   4   4   3  \r
0000020  \n   0   1   0   1   3   6       7       4   2  \r  \n   0   1
0000040   2   1   4   5       1   6       1  \r  \n   0   2   1   2   4
0000060   4       7   2       3  \r  \n   0   2   2   1   4   5       7
0000100   1       5   5
0000104

the values are different because i am using the example file to check outputs against the example outputs but should still act in the same way

example file and outputs

010134 756 443
010136 7 42
012145 16 1
021244 72 3
022145 71 55

outputs:
avg comp = 4141/5
avg wait = 3149/5

Answer 1

Assumptions/understandings (from OP's comments):

• input file has 3 columns: index time, burst time and id number
• each line of input contains details for one of the N sequentially run processes (where N is the number of lines in the input file)
• desired output are strings of the form SUM/N where SUM is the sum of execution/wait burst times for the N processes (aka number of lines in file)
• additional output is a string of the form ID1>ID2>ID3>...>IDN which is simply a concatenation of the values in the id number column
• OP has stated these outputs will not be used in follow-on 'math operations'
• it sounds (to me) like the whole purpose of this exercise is to generate some textual output representing aggregate metadata (of the inputs)

One bash idea:

sum=0
count=0

comp_total=0
wait_total=0

exec_order=
pfx=

while read -r index burst id
do
    exec_order+="${pfx}${id}"
    pfx='>'

    (( count++ ))

    wait_total=$((wait_total + sum))
    sum=$((sum + burst))
    comp_total=$((comp_total + sum))

done < Process.dat

#echo "comp_total : ${comp_total}"
#echo "wait_total : ${wait_total}"
    
comp_avg="${comp_total}/${count}"
wait_avg="${wait_total}/${count}"

echo "exec_order : ${exec_order}"
echo "comp_avg : ${comp_avg}"
echo "wait_avg : ${wait_avg}"

This outputs:

exec_order : 4433>54>534>12>73
comp_avg : 6132/5
wait_avg : 4174/5