How do I use awk/sed to merge a field across multiple rows based on matching column values?

Question

I am working with a CSV in bash, and attempting to merge the data in the 2nd column by matched data in the 3rd column.

My code works but the information in the other columns ends up just getting repeated instead of properly copied.

awk -F',' -v OFS=',' '{
            env_name=$1
            app_name=$4
            lob_name=$5
            if ($3 in a) {
                a[$3] = a[$3]" "$2;
            } else {
                a[$3] = $2;
            }
        }
        END { for (i in a) print env_name, i, a[i], app_name, lob_name}' input.tmp > output.tmp

This:

A,1,B,C,D
A,2,B,C,D
A,3,E,F,G
A,4,X,Y,Z
A,5,E,F,G

Should become this:

A,1 2,B,C,D
A,3 5,E,F,G
A,4,X,Y,Z

But instead we are getting this:

A,1 2,B,C,D
A,3 5,E,C,D
A,4,X,C,D

Answer 1

your grouping key should be all except second field

$ awk -F, 'BEGIN {SUPSEP=OFS=FS} 
                 {k=$1 FS $3 FS $4 FS $5; a[k]=(k in a)?a[k]" "$2:$2} 
           END   {for(k in a) {split(k,p); print p[1],a[k],p[2],p[3],p[4]}}' file

A,1 2,B,C,D
A,3 5,E,F,G
A,4,X,Y,Z

perhaps can be simplified a bit

$ awk 'BEGIN {OFS=FS=","} 
             {v=$2; $2=""; k=$0; a[k]=(k in a?a[k]" "v:v)}
       END   {for(k in a) {$0=k; $2=a[k]; print}}' file

Answer 2

sed + sort + awk

$ sed 's/,/+/3;s/,/+/3' merge_csv | sort -t, -k3 | awk -F, -v OFS=, ' { if($3==p) { a=a b " "; } if(p!=$3 && NR>1) { print $1,a b,p; a="" } b=$2; p=$3 } END { print $1,a b,p } ' | tr '+' ','
A,1 2,B,C,D
A,3 5,E,F,G
A,4,X,Y,Z

$

If Perl is an option, you can try this

$ perl -F, -lane '$x=join(",",@F[-3,-2,-1]); @t=@{$kv{$x}};push(@t,$F[1]);$kv{$x}=[@t]; END { for(keys %kv) { print "A,",join(" ",@{$kv{$_}}),",$_" }} ' merge_csv
A,1 2,B,C,D
A,4,X,Y,Z
A,3 5,E,F,G

$

Input file:

$ cat merge_csv
A,1,B,C,D
A,2,B,C,D
A,3,E,F,G
A,4,X,Y,Z
A,5,E,F,G

$