Another easy Bourne Shell Scripting question

WinterSt · 09-02-2004, 03:07 AM

It is very simple this time (but not simple enough for me). I have file with data entries seperated by ",".
The files looks like this..

Code:

Anderson,Sarah,02-95437869 
Brown,Sarah,02-99893878 
Couch,David,02-95673456 
Jones,Sarah,02-95671660 
Smith,John,02-93272658 
Williams,Nick,02-98781987

I want to display it with each field placed into left justified columns.

Code:

Anderson(white space)Sarah(white space)02-95437869 
Brown(white space''')Sarah(white space)02-99893878 
Couch(white space''')David(white space)02-95673456 
Jones(white space''')Sarah(white space)02-95671660 
Smith(white space''')John(white space')02-93272658 
Williams(white space)Nick(white space')02-98781987

I thought that I would be able to do this by adding multiple \t between each field, but the command I thought would do it, doesn't work.
cat records.txt | tr "[,]" "[\t*2]"

It only replaces the "," with one "\t".

What could I try to do to display this data in columns, or perhaps a way to add multiple tabs?

P.S has to work in bourne shell (#!/bin/sh).

thanksgsg

Tinkster · 09-02-2004, 03:50 AM

Another easy homework-question? :)

itsme86 · 09-02-2004, 03:59 AM

This one works for me:

Code:

#!/bin/sh

cat records.txt | sed 's/,/        /g'

That big space in there is where you shoud press the TAB key however many times you want.

jlliagre · 09-02-2004, 04:09 AM

However, it doesn't aligns well when fields have too distinct width.

Here's a try:

Code:

#!/bin/sh
awk -F, '
function format(s)
{
        padding="";
        l=length(s);
        while(l<22)
        {
                padding=padding " ";
                l++;
        }
        return(s padding);
}
{
        printf("%s %s %s\n",format($1),format($2),$3);
}' $1

Tinkster · 09-02-2004, 04:17 AM

Quote:

#!/bin/sh

cat records.txt | sed 's/,/ /g'

That's ugly ;) and uses cat in excess ... how about?

Code:

awk -F, '{printf("%-10s%-10s%-10s\n", $1,$2,$3)}' records.txt

And if you make the 10's just one higher than the length
of the longest string in each column you don't have to
worry about potential breaks of the structure ...

Or this very simple approach :)

Code:

column -t -s, records.txt

Cheers,
Tink

jlliagre · 09-02-2004, 11:38 AM

You win !

tonyfreeman · 09-02-2004, 01:47 PM

Wow ... that column command is bloody brilliant!

WinterSt · 09-02-2004, 07:41 PM

Thank you everyone, and especially Tinkster, my bourne shell does not have the column program, but the awk command you listed worked perfectly.

Thanks!