How can I split comma separated values into multiple rows? - bash

Question

I'm trying to split multiple comma-separated values into rows.
I have achieved it in a small number of columns with comma-separated values (with awk), but in my real table, I have to do it in 80 columns. So, I'm looking for a way to iterate.
An example of input that I need to split:
CHROM POS REF ALT GT_00 C_00 D_OO E_00 F_00 GT_11
chr1 10 T A 2,2 1,1 0,1 1,2 1,0 2
chr1 10 T G 3 2 1 2 0 0
The expected output:
chr1 10 T A 2 1 0 1 1 2
chr1 10 T A 2 1 1 2 0 2
chr1 10 T G 3 2 1 2 0 0
I have done it with the following code:
awk 'BEGIN{FS=OFS="\t"}
{
j=split($5,a,",");split($6,b,",");
split($7,c,",");split($8,d,",");split($9,e,",");
for(i=1;i<=j;++i)
{
$5=a[i];$6=b[i];$7=c[i];$8=d[i];$9=e[i];print
}}'
But, as I have said before, there are 80 columns (or more) with comma-separated values in my real data.
Is there a way to it using iteration?
Note: I need to do it in bash (not MySQL, SQL, python...)

Answer 1

This awk may do:
file:
chr1 10 T A 2,2 1,1 0,1 1,2 1,0 2
chr1 10 T G 3 2 1 2 0 0
chr1 10 T C 5 1,2,3 4,2,1 7 1,8,3 3
chr1 10 T D 1,2,3,5 4,2,1,8 1,8,3,2 3 5 7
Solution:
awk '{
n=0;
for(i=5;i<=NF;i++) {
t=split($i,a,",");if(t>n) n=t};
for(j=1;j<=n;j++) {
printf "%s\t%s\t%s\t%s",$1,$2,$3,$4;
for(i=5;i<=NF;i++) {
split($i,a,",");printf "\t%s",(a[j]?a[j]:a[1])
};
print ""
}
}' file
chr1 10 T A 2 1 0 1 1 2
chr1 10 T A 2 1 1 2 1 2
chr1 10 T G 3 2 1 2 0 0
chr1 10 T C 5 1 4 7 1 3
chr1 10 T C 5 2 2 7 8 3
chr1 10 T C 5 3 1 7 3 3
chr1 10 T D 1 4 1 3 5 7
chr1 10 T D 2 2 8 3 5 7
chr1 10 T D 3 1 3 3 5 7
chr1 10 T D 5 8 2 3 5 7
Your test input gives:
chr1 10 T A 2 1 0 1 1 2
chr1 10 T A 2 1 1 2 1 2
chr1 10 T G 3 2 1 2 0 0
It does not mater if comma separated values are consecutive, as long as you do not mix 2 or 3 comma on the same line.

Answer 2

Here is another awk. In contrast to the previous solutions where we split fields into arrays, we attack the problem differently using substitutions. There is no field iterating going on:
awk '
BEGIN { OFS="\t" }
{ $1=$1;t=$0; }
{ while(index($0,",")) {
gsub(/,[[:alnum:],]*/,""); print;
$0=t; gsub(OFS "[[:alnum:]]*,",OFS); t=$0;
}
print t
}' file
how does it work:
The idea is based on 2 types of substitutions:
gsub(/,[[:alnum:],]*/,""): this removes all substrings made from alphanumeric characters and commas that start with a comma: 1,2,3,4 -> 1. This does not change fields that have no comma.
gsub(OFS "[[:alnum:]]*,",OFS): this removes alphanumeric characters followed by a single comma and are in the beginning of the field: 1,2,3,4 -> 2,3,4
So using these two substitutions, we iterate until no comma is left. See How can you tell which characters are in which character classes? on details for [[:alnum:]]
input:
chr1 10 T A 2,2 1,1 0,1 1,2 1,0 2
chr1 10 T G 3 2 1 2 0 0
chr1 10 T C 5 1,2,3 4,2,1 7 1,8,3 3
chr1 10 T D 1,2,3,5 4,2,1,8 1,8,3,2 3 5 7
output:
chr1 10 T A 2 1 0 1 1 2
chr1 10 T A 2 1 1 2 0 2
chr1 10 T G 3 2 1 2 0 0
chr1 10 T C 5 1 4 7 1 3
chr1 10 T C 5 2 2 7 8 3
chr1 10 T C 5 3 1 7 3 3
chr1 10 T D 1 4 1 3 5 7
chr1 10 T D 2 2 8 3 5 7
chr1 10 T D 3 1 3 3 5 7
chr1 10 T D 5 8 2 3 5 7