Replace first 10 spaces in text file

[Robert A. Pierce]

Good morning!

I have a file like this:

a bb ccc dd ee ff gg hh iii j kk lmn op qr st

and I'd like to make it look like this:

a,bb,ccc,dd,ee,ff,gg,hh,iii,j,kk lmn op qr st

I figure I can use awk with

print $1","$2","$3...$12,$13...$NF

and so forth, but I'm not sure how to get from $11 to $NF. Perhaps something like $NF - $NF % 10, $NF + 1 - $NF % 10, and loop until $NF is reached?

Now I'm using "," and printing out to $20, which _should_ be greater than the number of fields, but since I don't have control over the data, I can't be sure of it staying that way.

This part doesn't have to be in awk; sed or unobfuscated perl would work as well.

Thanks,

Rob

[Ken Penland_1 · Accepted solution]

well, I am sure someone could come up with a shorter way to do it, but here is my solution to the problem, via perl where "textfile" is the file you are looking through:

#!/usr/local/bin/perl
foreach $line(`cat textfile`)
{
@bits = split(/\s+/,$line);
for ($i=0;$i < $#bits;$i++)
{
if ($i < 10) { print "$bits[$i],"; }
else { print "$bits[$i] "; }
}
print "\n";
}

[A. Clay Stephenson]

I would split $0 into an array and then your task becomes rather simple:

Here's x.awk:

{
i = 1
n = split($0,arry1," ")
while (i <= 10 && i < n)
{
printf("%s,",arry1[i])
++i
}
while (i < n)
{
printf("%s ",arry1[i])
++i
}
printf("%s\n",arry1[i])
}

awk -f x.awk < infile > outfile

[Robert A. Pierce]

Ken,

> I am sure someone could come up with a
> shorter way to do it

Probably, but yours works and is easy to read! Thanks a lot.

Clay,

> I would split $0 into an array and then
> your task becomes rather simple:

Ah, yes! The array makes it very simple. Thanks very much!

Rob

[Chris Vail]

You could convert all the spaces in your document to commas with
cat file|tr -s " " ","

then handle it with awk thereafter.



Chris

[TwoProc]

A quick n goofy way...
Capture the whole list in a variable...
Set the list as your environment variables
This will put the first one in $1
The second one in $2
And the number of them in $#
That's All you need to know...

#!/bin/ksh
A="aa bb cc dd ee "
set $A
integer i; i=$#
while [ $i -gt 0 ]
do
echo $i
i=${i}-1
done

output...
aa
bb
cc
dd
ee

[H.Merijn Brand (procura]

Just for the purists amongs you, the shorter way is to use -a

# echo "a bb ccc dd ee ff gg hh iii j kk lmn op qr st" | perl -nlae'splice@F,0,10,join",",@F[0..9];print join" ",@F'
a,bb,ccc,dd,ee,ff,gg,hh,iii,j kk lmn op qr st
#

from 'man pelrun':
--8<---
-a turns on autosplit mode when used with a -n or -p.
An implicit split command to the @F array is done as
the first thing inside the implicit while loop pro�
duced by the -n or -p.

perl -ane 'print pop(@F), "\n";'

is equivalent to

while (<>) {
@F = split(' ');
print pop(@F), "\n";
}

An alternate delimiter may be specified using -F.
-->8---

>> perl -nlae

-n: the -e section is a non-print loop around each line
-l: print a newline after every print (not printf) statement
-a: autosplit on whitespace to @F
-e: code section

>> splice@F,0,10,join",",@F[0..9];

replace, starting with the first element (index 0), 10 elements of @F, with the join with space of the first 10 elements (0..9) of @F

>> print join" ",@F

print all fields joined with space

Now play some golf (achief the same in less key-strokes):

perl -plae'splice@F,0,10,join",",@F[0..9];$_=join" ",@F'

or even

perl -plae'$_=qq{@{[(join",",@F[0..9]),@F[10..$#F]]}}'

and my final offer

perl -plae'$_=join" ",join(",",@F[0..9]),@F[10..$#F]'

Enjoy, Have FUN! H.Merijn [ who still thinks perl is so much fu

[Robert A. Pierce]

Chris,

The `tr` program is great, except this particular data file has some variable-length text fields at the end; some of these fields may contain spaces.

What I needed to do was modify the _first_ part from "space-delimited" to "comma or pipe-delimited" data, and then I could work with the rest.

So, the problem was 1) space-delimited data with spacey text-fields on the end, and 2) a variable number of fields (so the original awk print $2","$3..$12,$13 wouldn't work).

(Okay, the primary problem was inflexible output to start with, but I couldn't control that. :)

Breaking the lines into arrays (in Perl and in Awk) fixed it.

[Robert A. Pierce]

John,

> A quick n goofy way...

Thanks! That's pretty neat.

Merijn,

> the shorter way

With wonderful examples and explanations!

Vele dank!