Data::Record(3) split on steroids

VERSION

Version 0.02

SYNOPSIS


use Regexp::Common;
use Data::Record;
my $record = Data::Record->new({
split => "\n",
unless => $RE{quoted},
});
my @data = $record->records($data);

DESCRIPTION

Sometimes we need data split into records and a simple split on the input record separator ($/) or some other value fails because the values we're splitting on may allowed in other parts of the data. Perhaps they're quoted. Perhaps they're embedded in other data which should not be split up.

This module allows you to specify what you wish to split the data on, but also speficy an ``unless'' regular expression. If the text in question matches the ``unless'' regex, it will not be split there. This allows us to do things like split on newlines unless newlines are embedded in quotes.

METHODS

new

Common usage:

 my $record = Data::Record->new({
    split  => qr/$split/,
    unless => qr/$unless/,
 });

Advanced usage:

 my $record = Data::Record->new({
    split  => qr/$split/,
    unless => qr/$unless/,  # optional
    token  => $token,       # optional
    chomp  => 0,            # optional
    limit  => $limit,       # optional (do not use with trim)
    trim   => 1,            # optional (do not use with limit)
    fields => {
        split  => ',',
        unless => $RE{quoted}, # from Regexp::Common
    }
 });

The constructor takes a hashref of key/value pairs to set the behavior of data records to be created.

  • split

    This is the value to split the data on. It may be either a regular expression or a string.

    Defaults to the current input record separator ($/).

  • unless

    Data will be split into records matching the split value unless they also match this value. No default.

    If you do not have an "unless" value, use of this module is overkill.

  • token

    You will probably never need to set this value.

    Internally, this module attempts to find a token which does not match any text found in the data to be split and also does not match the split value. This is necessary because we mask the data we don't want to split using this token. This allows us to split the resulting text.

    In the unlikely event that the module cannot find a token which is not in the text, you may set the token value yourself to some string value. Do not set it to a regular expression.

  • chomp

    By default, the split value is discarded (chomped) from each record. Set this to a true value to keep the split value on each record. This differs slightly from how it's done with split and capturing parentheses:

      split /(\,)/, '3,4,5';
    

    Ordinarily, this results in the following list:

     ( 3, ',', 4, ',', 5 )
    

    This module assumes you want those values with the preceding record. By setting chomp to false, you get the following list:

     ( '3,', '4,' 5 )
    
  • limit

    The default split behavior is similar to this:

     split $split_regex, $data;
    

    Setting "limit" will cause the behavior to act like this:

     split $split_regex, $data, $limit
    

    See "perldoc -f split" for more information about the behavior of "limit".

    You may not set both "limit" and "trim" in the constructor.

  • trim

    By default, we return all records. This means that due to the nature of split and how we're doing things, we sometimes get a trailing null record. However, setting this value causes the module to behave as if we had done this:

     split $split_regex, $data, 0;
    

    When "split" is called with a zero as the third argument, trailing null values are discarded. See "perldoc -f split" for more information.

    You may not set both "limit" and "trim" in the constructor.

    Note: This does not trim white space around returned records.

  • fields

    By default, individual records are returned as strings. If you set "fields", you pass in a hashref of arguments that are identical to what "new" would take and resulting records are returned as array references processed by a new "Data::Record" instance.

    Example: a quick CSV parser which assumes that commas and newlines may both be in quotes:

     # four lines, but there are only three records! (newline in quotes)
     $data = <<'END_DATA';
     1,2,"programmer, perl",4,5
     1,2,"programmer,
     perl",4,5
     1,2,3,4,5
     END_DATA
      
     $record = $RECORD->new({
         split  => "\n",
         unless => $quoted,
         trim   => 1,
         fields => {
             split  => ",",
             unless => $quoted,
         }
     });
     my @records = $record->records($data);
     foreach my $fields (@records) {
       foreach my $field = (@$fields);
         # do something
       }
     }
    

    Note that above example will not remove the quotes from individual fields.

split

  my $split = $record->split;
  $record->split($on_value);

Getter/setter for split value. May be a regular expression or a scalar value.

unless

 my $unless = $self->unless;
 $self->unless($is_value);

Getter/setter for unless value. May be a regular expression or a scalar value.

chomp

  my $chomp = $record->chomp;
  $record->chomp(0);

Getter/setter for boolean chomp value.

limit

  my $limit = $record->limit;
  $record->limit(3);

Getter/setter for integer limit value.

trim

  my $trim = $record->trim;
  $record->trim(1);

Getter/setter for boolean limit value. Setting this value will cause any previous "limit" value to be overwritten.

token

  my $token = $record->token;
  $record->token($string_not_found_in_text);

Getter/setter for token value. Token must be a string that does not match the split value and is not found in the text.

You can return the current token value if you have set it in your code. If you rely on this module to create a token (this is the normal behavior), it is not available via this method until "records" is called.

Setting the token to an undefined value causes Data::Record to try and find a token itself.

If the token matches the split value, this method will croak when you attempt to set the token.

If the token is found in the data, the "records" method will croak when it is called.

records

  my @records = $record->records($data);

Returns @records for $data based upon current split criteria.

BUGS

It's possible to get erroneous results if the split value is "/\d+/". I've tried to work around this. Please let me know if there is a problem.

CAVEATS

This module must read all of the data at once. This can make it slow for larger data sets.

AUTHOR

Curtis ``Ovid'' Poe, "<ovid [at] cpan [dot] org>"

BUGS

Please report any bugs or feature requests to "[email protected]", or through the web interface at <http://rt.cpan.org/NoAuth/ReportBug.html?Queue=Data-Record>. I will be notified, and then you'll automatically be notified of progress on your bug as I make changes.

ACKNOWLEDGEMENTS

Thanks to the Monks for inspiration from <http://perlmonks.org/index.pl?node_id=492002>.

0.02 Thanks to Smylers and Stefano Rodighiero for catching POD errors.

COPYRIGHT & LICENSE

Copyright 2005 Curtis ``Ovid'' Poe, all rights reserved.

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.