Column Read and Average (BASH) - bash

Question

so I'm dealing with column averaging problems
I have several columns of number,
1 2 3 4 5
2 3 4 5 6
8 4 5 6 7
9 8 7 6 5
1 9 9 9 2
What I want to do is to take an average of each column so that they will yield
4 5 5 6 5
The summing part is not a problem, I'm able to get the sum of each column. However, somehow I'm having trouble in counting the number of columns (variable $x)
Here is my code
while read line; do
x = 0
read -a array <<< "$line"
for i in "${!array[#]}"
do
column[${i}]=$((${column[${i}]}+${array[$i]}))
((x++))
done
done < $TMP
for sum in ${column[#]}
do
average=`expr $sum / $x`
remainder=`expr $sum % $x`
mult=`expr $remainder \* 10`
fracvalue=`expr $mult / $x`
echo $x
echo $average
echo $sum
echo $remainder
echo $mult
echo $fracvalue
done
The last several lines are for my testing purposes, here the $X keeps showing 1 instead of 5. That's why it screws up all the other variable. Does anybody know where's the flaw in this code? Really appreciate your help. Thanks

Answer 1

Another approach:
#!/bin/bash
declare -ia sum # declare array of integers
declare -i lines=0 # declare integer to count lines
# read file to array and sum values
while read -ra array; do
for ((i=0; i<${#array[#]};i++)); do
sum[$i]+=${array[$i]}
done
lines+=1
done < file
# print averages
for ((j=0; j<$i;j++)); do
echo -n "$((${sum[$j]}/$lines)) "
done
Output:
4 5 5 6 5

Answer 2

The problem is in the line
x = 0
should be
x=0 and make ((x+1)) as let "x+=1" it should work.

Answer 3

Maybe your file doesn't have the same number of lines in each column.
while read line; do
read -a array <<< "$line"
for i in "${!array[#]}"
do
column[${i}]=$((${column[${i}]}+${array[$i]}))
((x[${i}]++))
done
done < $TMP
for i in ${!column[#]}
do
sum=${column[$i]}
x=${x[${i}]}
#your calcul
done

Answer 4

The initial issue, aside from a few syntax peculiarities, was the resetting of x on each read which caused you to lose your line count. Additionally, your array is an indexed array, not an associative array.
Making the small adjustments and declareing the variables at the beginning (to hint to bash they are integers or arrays), it works as planned. note: there is no need for a here-string to assign the array, just use regular array syntax. Also note, for indexed arrays, there is no need to dereference the variable within [ ], simply ${array[i]} is fine.
While the use of expr is fine (old, slow, but portable and fine), you could use bash arithmetic syntax for your parameter calculations at the end. (you already have arrays). Finally, you are better served initializing each column[$i]=0 on the first read loop. You can do that with a simple flag:
#!/bin/bash
declare -i ncols=0
declare -i x=0
declare -a column
while read -r line; do
array=( $line )
[ $ncols -eq 0 ] && ncols=${#array[#]}
for ((i = 0; i < ncols; i++))
do
[ $x -eq 0 ] && column[$i]=0
column[$i]=$(( ${column[i]} + ${array[i]}))
done
((x++))
done < "$1"
for sum in ${column[#]}
do
average=`expr $sum / $x`
remainder=`expr $sum % $x`
mult=`expr $remainder \* 10`
fracvalue=`expr $mult / $x`
echo $x
echo $average
echo $sum
echo $remainder
echo $mult
echo $fracvalue
done
exit 0
Output
$ bash colavg.sh dat/colavg.txt
5
4
21
1
10
2
5
5
26
1
10
2
5
5
28
3
30
6
5
6
30
0
0
0
5
5
25
0
0
0