How to multiply columns in Bash? - bash

Question

I have the following three column data (first row is header) in a csv format
Value
Y
X
A
8
2
B
3
5
C
7
9
I want the following output also in a csv format
Value
Y*X
AB
40
BA
6
AC
72
CA
14
BC
27
CB
35
Is there a way to accomplish this in bash?
thank you
Here is the csv file copy-paste
VALUE,Y,X
A,13,7
C,0,0
D,3,25
E,2,44
F,0,6
H,1,1
I,5,3
K,45,3
L,1,31
M,2,3
N,3,3
P,113,87
Q,13,11
R,20,5
S,7,9
T,9,4
V,7,3
Y,1,0
I tried awk '{print $2*$3}' TEST.dat but the problem is it is not combinatorial i.e. not multiplying every column 2 value to every column 3 value.

Answer 1

Using awk:
awk -F, 'BEGIN { print "VALUE,Y*X"; i=0 } # Print header
FNR == 1 { next } # Skip existing header lines
FNR == NR { x[++i]=$3; values[i]=$1; next } # First pass through the file
{ # Second pass; multiply current row against every saved row
for (n = 1; n <= i; n++)
if (values[n] != $1) # Except itself
printf "%s%s,%d\n", $1, values[n], $2 * x[n]
}' input.csv input.csv
Process the file twice; first time saving the x values, and second time multiplying the current line's y against all the saved x's.
For fun, a version that uses sqlite, importing the CSV file and then doing a self-join:
sqlite3 -batch -header -csv <<EOF
.import input.csv data
SELECT a.value || b.value AS "VALUE", a.y * b.x AS "Y*X"
FROM data AS a
JOIN data AS b ON a.value <> b.value
ORDER BY a.rowid, b.rowid;
EOF
And pure bash:
#!/usr/bin/env bash
declare -a values yvalues xvalues
exec 3<input.csv
read -r -u 3 _ # Read and discard header
declare -i i=0
while IFS=, read -r -u 3 value y x; do
i+=1
values[i]=$value
yvalues[i]=$y
xvalues[i]=$x
done
echo "VALUE,Y*X"
for ((a=1; a<=i; a++)); do
for ((b=1; b<=i; b++)); do
if [[ $a -ne $b ]]; then
printf "%s%s,%d\n" "${values[a]}" "${values[b]}" \
"$(( yvalues[a] * xvalues[b] ))"
fi
done
done