Question

I have something like a csv file in which the field delimiter is “@”.

ID@Name@Surname@Age@Profession@Address

1254343123@John@Smith@24@Engineer@Washington
23@Alexander@Kristofferson-Brown@Economic Advisor@Kent
...

I want to convert it into something more human-readable as in:

1254343123    John        Smith                  24    Engineer            Washington
23            Alexander   Kristofferson-Brown    35    Economic Advisor    Kent

…where each column stars at a specific character.

I have tried something that takes advantage of the TAB-size value and adds several TABs to the field:

sed -e "{/@[^@]\{32,\}@/s/@\([^@]*\)@/\t\1\t/g};{/@[^@]\{24,31\}@/s/@\([^@]*\)@/\t\1\t/g};{/@[^@]\{16,23\}@/s/@\([^@]*\)@/\t\1\t\t/g};{/@[^@]\{8,15\}@/s/@\([^@]*\)@/\t\1\t\t/g};{/@[^@]\{2,7\}@/s/@\([^@]*\)@/\t\1\t\t\t/g}"

…which does not work in all cases.

Could somebody give me a hint as how to proceed?

PS: My intention is to use mainly sed (a one-liner) or if sed just doesn’t cut it, awk is OK.

Answer 1

awk -F@ '{for(i=1;i<=NF;i++){printf "%-20s", $i};printf "\n"}' input.csv

Input

$ cat input.csv
1254343123@John@Smith@24@Engineer@Washington
23@Alexander@Kristofferson-Brown@35@Economic Advisor@Kent

Output

$ awk -F@ '{for(i=1;i<=NF;i++){printf "%-20s", $i};printf "\n"}' input.csv
1254343123          John                Smith               24                  Engineer            Washington
23                  Alexander           Kristofferson-Brown 35                  Economic Advisor    Kent

If you want to make the field width (20 in the code above) a shell variable that can be passed in you do something like this:

#!/bin/bash

fldwth=20

awk -v fw=$fldwth -F@ '{for(i=1;i<=NF;i++){printf "%-*s", fw,$i};printf "\n"}' input.csv