Removing "," (comma) value from the columns of a CSV file using linux shell

I'm having a CSV file with 5 columns where 1 column is having integer values like (1,564, 4,789, 9,765). I need to remove "," (comma) from col3 where integer value should look like as follows: (1564,4789,9865).

Tried various ways using "awf" & "sed"
but none of it is working.
Below is the sample input and expected output.
Input CSV file:

col1,col2,col3,col4,col5
Abcd,defg,"1,432",hjik,khuj

Output CSV file:

col1,col2,col3,col4,col5
Abcd,defg,"1432",hjik,khuj

This regex might help you find the unquoted fields:

(\d{1,3}(?:,\d{3})+)

and with a simple replacement, like "\1", properly quote them.

It looks for "1-3 digits, followed by some number of sequences of a comma and three digits".

I say might because you've only shown one small sample of the data, and this only works on numbers sensibly separated with commas.

I started with this broken CSV:

col1,col2,col3
a,1,b
a,12,b
a,123,b
a,1,234,b
a,12,345,b
a,123,456,b
a,1,234,567,b

and got this:

| c1 | c2        | c3 |
|----|-----------|----|
| a  | 1         | b  |
| a  | 12        | b  |
| a  | 123       | b  |
| a  | 1,234     | b  |
| a  | 12,345    | b  |
| a  | 123,456   | b  |
| a  | 1,234,567 | b  |

You can see it in action here, regexr.com/7ht81:

This small Python program will fix my sample CSV:

import re

lines: list[str] = []
with open("input.csv", encoding="utf-8") as f:
    for line in f:
        fixed = re.sub(r"(\d{1,3}(?:,\d{3})+)", '"\"', line)
        lines.append(fixed)

with open("output.csv", "w", encoding="utf-8") as f:
    f.writelines(lines)

Fist replace the commas you want to keep with \r, next remove the others.
Finally restore the commas.

sed -r 's/([^,]*),([^,]*),(.*),([^,]*),([^,]*)/\r\r\r\r/;s/,//g;s/\r/,/g' input.csv

EDIT:
As Zach commented, you might want to replace all \r with |||||. On Linux you expect only \n characters, but a csv format is often generated on Windows.

You should use a CSV aware tool as Miller.

Running

mlr --csv put '$col3=gsub($col3,",","")' input.csv >output.csv

you get

col1,col2,col3,col4,col5
Abcd,defg,1432,hjik,khuj

• put is the verb to apply functions.

• gsub is the function that substitutes , with nothing in col3

use below to replace comma with space,dot or anything

#!/bin/sh

read filename

#replace all_occurrence(g) of  comma with .
sed 's/,/./g' "$filename" >> output.csv