Prev: Data cleaning issue involving bad wide characters in what ought to be ascii data
Next: SIGUSR1 ignored during poll() if sleep() and warn() called?
From: toralf on 9 Sep 2009 10:39
I've a file containing tab separated values - most, but not all are
quoted - and now I'm wondering how to substitute a non-quoted value like
<tab>20090807<tab> by sth. like <tab>"20090807"<tab>

--
MfG/Sincerely

Toralf F�rster
pgp finger print: 7B1A 07F4 EC82 0F90 D4C2 8936 872A E508 7DB6 9DA3


From: David Harmon on 9 Sep 2009 12:43
On Wed, 9 Sep 2009 14:39:59 +0000 (UTC) in comp.lang.perl.misc, toralf
<toralf.foerster(a)gmx.de> wrote,
>I've a file containing tab separated values - most, but not all are
>quoted - and now I'm wondering how to substitute a non-quoted value like
><tab>20090807<tab> by sth. like <tab>"20090807"<tab>

What part are you having trouble with? What did you try and what
happened?

I suggest, use regex substitution s/// with \t for tab, capture the
string of digits, then use $1 to put them back in the replacement
string. Don't forget to account for the first and last fields not
having both tabs.

From: sln on 9 Sep 2009 15:09
On Wed, 9 Sep 2009 14:39:59 +0000 (UTC), toralf <toralf.foerster(a)gmx.de> wrote:

>I've a file containing tab separated values - most, but not all are
>quoted - and now I'm wondering how to substitute a non-quoted value like
><tab>20090807<tab> by sth. like <tab>"20090807"<tab>

You would have to do some complicated regex (possibly multiple regx's),
but that depends on the data sample in regards to quote's, newlines, or other
shapes it can have.
A simple way is to split on tab, fix up the value, the join it back together.
Its ugly though.

-sln
------------
use strict;
use warnings;

my @ar = (
qq{\n12345145\t\n36367\t"qfqqbv"\n\t"0987"\t"asdf"a"\n },
qq{\t"01234"\taaaa\t494848\t}
);

for my $str (@ar) {
my $newstring = join "<tab>", map {/^(\s*|)"*(.*?|)"*(\s*|)$/; $1.'"'.$2.'"'.$3 } split (/\t/,$str);
print "-> $newstring\n\n";
}
__END__

From: David Harmon on 9 Sep 2009 20:19
On Wed, 09 Sep 2009 12:09:05 -0700 in comp.lang.perl.misc,
sln(a)netherlands.com wrote,

> (\s*|)

Just in case \s* fails to match the empty string, eh?

From: sln on 9 Sep 2009 22:09
On Wed, 09 Sep 2009 17:19:50 -0700, David Harmon <source(a)netcom.com> wrote:

>On Wed, 09 Sep 2009 12:09:05 -0700 in comp.lang.perl.misc,
>sln(a)netherlands.com wrote,
>
>> (\s*|)
>
>Just in case \s* fails to match the empty string, eh?

Doh, what was I thinking.
/^(\s*)"*(.*?)"*(\s*)$/

-sln
 |  Next  |  Last
Pages: 1 2
Prev: Data cleaning issue involving bad wide characters in what ought to be ascii data
Next: SIGUSR1 ignored during poll() if sleep() and warn() called?