removing specific character from specific column in unix file

I have unix file with csv format like below

   col1,  col2,  col3,  col4
   abc,   efg,   h,i,j  lmn

Need to remove "," from col3 and keep , separated values. Any help ?

>Solution :

Instead of using sed I suggest using awk for this. Since the columns are separated by one or more whitespaces, you can use the global substitution function gsub on the third column.

This replaces , (the regex /,/) with an empty string ("") in the third column ($3):

awk '{gsub(/,/,"",$3)}1' the_file