in reply to identifying null fields in bar delimited records

Something along the lines of 
my $line = ";field;;field;\t;field; ;field;field";
my @line = split ";", $line;
my @null = grep { defined $_ } map { $line[$_] =~ /^$/ ? $_ : undef } 
+(0..$#line);

print "line contains ", scalar @null, " null fields at offset(s): @nul
+l";
#line contains 2 null fields at offset(s): 0 2

[download]

?

Update: To fit in the OP's specified requirements:

use strict;

while ( <DATA> )
{
    my @line = split /\|/, $_;
    my @null = grep { defined $_ } map { $line[$_] =~ /^$/ ? $_ : unde
+f } (0..$#line);
    print "line contains ", scalar @null, " null fields at offset(s): 
+@null\n";
}

#line contains 3 null fields at offset(s): 1 3 4
#line contains 2 null fields at offset(s): 2 4
#line contains 3 null fields at offset(s): 0 1 4
#line contains 2 null fields at offset(s): 1 2 4

__DATA__
first||third||
alpha|beta||delta|
||c|d|
one|||four|

[download]


holli, /regexed monk/

Replies are listed 'Best First'.
Re^2: identifying null fields in bar delimited records
by wfsp (Abbot) on May 31, 2005 at 12:07 UTC
    holli++

    Here's a humble for loop: 

    #!/usr/bin/perl

use strict;
use warnings;

my $line = "a||c|\t|  |d||e";
my @fields = split(/\|/, $line);
my @null;

for my $i (0..$#fields){
  push @null, $i unless $fields[$i];
}

print "null fields: ", scalar @null, "\n";
print "at field: ";
print ++$_, ", " for @null;

# output
# null fields: 2
# at field: 2, 7,

    [download]
[reply]
[d/l]
      holli and wfsp, great answers. Now, I'd like to make this snippet of code more general by turning the delimiter into a variable. But doing so doesn't work (I'm struggling to understand regular expressions). 
      my $delim = '|';
my $line = "a||c|\t|  |d||e";
my @fields = split(/\\$delim/, $line);

# output
# null fields: 0
# at field:

      [download]
[reply]
[d/l]
        my $delim = qr/\|/;
my $line = "a||c|\t|  |d||e";
my @fields = split($delim, $line);

        [download]

        The qr operator quotes and compiles its STRING as a regular expression.
        split splits on a pattern (regex).

        update:

        Added explanation.

[reply]
[d/l]
[select]
        I would put the escaping backslash into $delim, so you are truly independent. Besides that your code works fine and splits the string as intended. 
        my ($delim, $line, @fields);

$delim = '\|';
$line  = "a|c||b";
@fields = split(/$delim/, $line);
print join ("*", @fields), "\n";

$delim = ';';
$line  = "a;c;;b";
@fields = split(/$delim/, $line);
print join ("*", @fields), "\n";


#a*c**b
#a*c**b

        [download]
        What else did you expect?


        holli, /regexed monk/
[reply]
[d/l]
      I posted as anon by mistake, which means I can't edit my post to remove the superfluous line containing the variable "line".
[reply]
Re^2: identifying null fields in bar delimited records
by Anonymous Monk on May 31, 2005 at 14:01 UTC
    I ended up doing this in awk; it could probably be translated to perl by someone smarter than me. For some reason the filename is printed AFTER the results instead of before as I would have expected, but otherwise this seems to work.
    BEGIN{FS="\|";}
{
   line= substr($0,0,length($0)-1); #peel trailing bar
   n=split($0,a);
   for (i=0;i<n;i++){
      line=line  a[i];
      if(a[i] ==""){nulls[i]++};
   }
}
END{
   print FILENAME;
   for (i in nulls) print "\t field " i": " nulls[i] " nulls";
}

    [download]
[reply]
[d/l]

In Section Seekers of Perl Wisdom