Extract rows using a delimiter and append as column in a file

With perl

$ perl -lane '@v=split/,/,$F[1]; $i=0; print "$_ $v[$i++]" for split/,/,$F[0]' ip.txt a 1 b 2 c 3 d 4 e 5 f 6 g 7 h 8 

Split either the first/second column, initialize index counter and then print the pairs by iterating over splitting the other column

-a option will auto-split the input line on whitespaces and have the results in @F array

Awk solution (assuming that number of "keys" (contained in the 1st field $1) would always correspond to number of "values" (contained in the 2nd field $2)):

awk '$1 ~ /,/{ len = split($1, keys, ","); split($2, vals, ","); for (i = 1; i <= len; i++) print keys[i], vals[i]; next }1' file 

The output:

a 1 b 2 c 3 d 4 e 5 f 6 g 7 h 8 

One way to do it using the sed editor is:

sed -e ' s/,/\n/ s/\(\n.*[[:blank:]]\)\([^,]*\),/ \2\1/ P;D ' input.file 

Working:

• Clip the leading comma-separated element from the 2nd field.
• Then append this element in the 1st field's leading comma-separated element.
• Print the leading element of the 1st field and delete it after this.
• Repeat this procedure with what remains in the pattern space till it is empty.

Another method using Perl is:

perl -lane ' my($kref, $vref, %h) = map { [split /,/] } @F[0,1]; @h{@$kref} = @$vref; print "$_ $h{$_}" for @$kref; ' input.file 

Another way is shown here:

perl -lpe 'print "$1 $3" while s/^([^,]*),(.*\h)([^,]*),/$2/' input.file 

Working:

• Look at the regex this way: (Perl reads in a line at a time from the file) then:
  • ^([^,]*) shall pick the first-field's leading comma-separated element of the current line. This is stored in $1 variable.
  • (.*\h) shall preserve, for the next iteration of the while loop, the intermediate contents beginning from the second comma-separated element of the first field to the beginning of the second comma-separated element of the second field. This is stored in the $2 variable.
  • ([^,]*) shall pick the leading comma-separated element from the second field of the current line.This gets stored in the $3 variable.
  • Now "$1 $3" is printed out to STDOUT and the line is shrunk to $2. The while loop now performs the operation all over again on this edited line, which is $2 of the previous line,..... this repeats till the s/// succeeds. The failure comes when we run out of commas. At which point , what remains in the line, "c 5" is printed to STDOUT by the default behavior of perl in the -p mode.
• Pluck out the leading comma-separated elements from the first and second fields.
• Print those elements and also shrink the current record by removing.
• Loop over the current record while it's having 2 commas.
• Last pair is auto-printed due to the -p option of Perl.

perl -lane ' my($kref, $vref) = map { [split /,/] } @F; print shift @$kref, " ", shift @$vref while @$kref && @$vref; ' input.file 

Working:

• keys are stored in an array @$kref, corresponding values in @$vref. Note no hashes involved here.
• print the top of the arrays simultaneously and then remove the top ...rinse,repeat while both the arrays are non-empty.

Output:

a 1 b 2 c 3 d 4 e 5 f 6 g 7 h 8 

Using awk:

awk '{gsub(","," "); for(i=0;i<NF/2;i++) print $(i+1),$(i+1+NF/2)}' file 

After removing the comma from each line, the script loops through half of the parameters of each line to print the first fields together with the field from the second half of the line.

"using shell scripting" -- this is bash:

while read -r key value; do IFS=, read -ra keys <<<"$key" IFS=, read -ra vals <<<"$value" for ((i=0; i < ${#keys[@]}; i++)); do echo "${keys[i]} ${vals[i]}" done done <<END a 1 b 2 c,d,e 3,4,5 f 6 g,h 7,8 END 

With gnu sed

sed -E ':A;s/([^,]*),([^ ]*) ([^,]*),(.*)/\1 \3\n\2 \4/;tA' infile 

Sorry for the obfuscation...

perl -pe '1 while s/(.*),(.*\h)(.*),/$1 $3\n$2/' infile 

Each substitution extracts last pair:

a,b,c 1,2,3 → a,b 1,2 → a 1 c 3 b 2 c 3 

There are a lot of solutions here programmed in perl, awk, sed, bash, ... but no script-solution.

# /bin/bash f="yourFile" paste -d' ' <(cut -d' ' -f1 "$f" | tr , '\n') <(cut -d' ' -f2 "$f" | tr , '\n') 

or written differently

# /bin/bash f() { cut -d' ' -f"$1" yourFile | tr , '\n'; } paste -d' ' <(f 1) <(f 2)