On Thu, 2004-05-27 at 15:09, Jose Alves de Castro wrote:
> On Thu, 2004-05-27 at 06:14, Jim Witte wrote:
> > Given a file of words W such as 'cat dog at home ...' (or perhaps read
> > into an array, though that would be a very large array), and a set of
> > letters L (a string 'aoeuidhtns' - perhaps put into a array), how would
> > I write a program to extract all words in W whose letters are all in L?
> > I'm thinking of a program to generate a string of words for a
> > typing-tutor program.
>
> Hello, Jim.
>
> I'm sorry if this seems complicated, but it really isn't :-) This is a
> one-liner that will do precisely what you want (I hope :-) )
>
> perl -pe 'BEGIN{$r=qr/^[asdfghjkl]+$/}split/ /;$_=join" ",grep/$r/,@_'
> file
>
> This particular one-liner will search for all words having only letters
> from the second alphabetic keyboard row (of qwerty keyboards, at least).
>
>
> Just to reassure you'll understand what is going on, I'm going to break
> the one-liner in some more readable code:
>
> =============================
>
> #!/usr/bin/perl -pw
> use strict;
>
> my $r;
>
> BEGIN {
> $r = qr/^[asdfghjkl]+$/;
> }
>
> split / /;
>
> $_ = join " ", grep /$r/, @_;
>
> =============================
>
> ...if you can call that readable...
>
> Now lets go step by step :-)
>
> =============================
>
> #!/usr/bin/perl -pw
> # the -p switch will makes all input lines be printed (after being
> processed)
> use strict;
> # I hope you know what this is for
>
> my $r; # this will be our regular expression
>
> BEGIN {
> $r = qr/^[asdfghjkl]+$/;
> } # and on this BEGIN block we assign it to be a sequence containing
> characters from the mentioned row *only*
>
> split / /;
> # here we split the input line by spaces (and the result is put into @_)
>
> $_ = join " ", grep /$r/, @_;
> # and now we grep the elements that correspond to $r and join them with
> spaces
>
> # et voilá! the -p switch makes the result be printed
>
> =============================
>
> HTH :-)
That was simply neat. I had read in a perl book ' there is always a
shorter way in perl '. I think this proves it.
But to think of it there is one hitch
suppose my string is 'god'
Assume
$word = "good"
$r = qr/^[god]+$/
then $r would match $word. Can you think of a good work around ?
Thanks
Ram
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
