Re: Generic way for replacing whitespaces with newlines

From: Ed Morton (morton_at_lsupcaemnt.com)
Date: 09/12/04 

Date: Sun, 12 Sep 2004 11:10:07 -0500

Kenny McCormack wrote:
> In article <svadnZA05Pxy-dncRVn-pg@comcast.com>,
> Ed Morton <morton@lsupcaemnt.com> wrote:
>
>>
>>Kenny McCormack wrote:
>><snip>
>>
>>>c) Since modern versions of awk (i.e., gawk, tawk) have built-in sorting (*),
>>>you can save a command and a pipe:
>>>
>>>awk -vdom="$MYLOCALDOMAIN"" '/^(domain|search)/ && $0 !~ dom {for (i=2;
>>>i <= NF ; i++) {zoo[$i]}} END {for (i in zoo) print i}' /etc/resolv.conf | xargs
>>>
>>>(*) In the distributed version of gawk, you have to set the environment
>>>variable "WHINY_USERS" in order for this to work.
>>
>>Tell me you're kidding!
>>
>> Ed.
>
>
> Nope. No joke. Observe:
>
> % gawk 'BEGIN {split("the quick brown fox jumps over the lazy dog",T);for (i in T) x[T[i]];for (i in x) print i}'
> fox
> quick
> over
> brown
> the
> jumps
> lazy
> dog
> % setenv WHINY_USERS
> % gawk 'BEGIN {split("the quick brown fox jumps over the lazy dog",T);for (i in T) x[T[i]];for (i in x) print i}'
> brown
> dog
> fox
> jumps
> lazy
> over
> quick
> the
> %
>

Unbelievable. Since we're getting into specific gawk details, I'm
setting followups to comp.lang.awk as I'd like to learn a little more
about this and that variable isn't mentioned in the gawk documentation
(http://www.gnu.org/software/gawk/manual/gawk.html).

Do you know:

a) Why "WHINY_USERS" was picked as the variable name?
b) Where is it documented?
c) Why was a variable chosen to implement this extremely useful
functionality rather than some language constauct (argument or "in"
extension, or...)?
d) Is there a way to specify that the sorting should be numeric instead
of alphabetical?
e) Can anyone explain the blank line at the start of the output in the
second example below?

eg1:
$ gawk 'BEGIN {split("10 1 2 3",T);for (i in T) x[T[i]];for (i in x)
print i}'
1
10
2
3
-----------

eg2:
$ gawk 'BEGIN {split("10 1 20 3",T);for (i in T) x[T[i]];for (i in x)
print i}'

1
10
20
3
------------

The above was produced with gawk 3.1.3 under Cygwin on Windows XP.

Regards,

        Ed.

Relevant Pages