/ 
Geo-Postcodes-0.30
River stage one • 2 direct dependents • 2 total dependents
/ Geo::Postcodes::Selection

NAME

Geo::Postcodes::Selection - How to use the selection procedure/method of the Geo::Postcodes::* modules

INTRODUCTION

This file uses the fictious country Utopia (with country code 'U2'), and the rest of this document will refer to the non-existent module Geo::Postcodes::U2.

SELECTION

selection procedure

my @postcodes = Geo::Postcodes::U2::selection($field => $value);

This simple form will give a list of postcodes matching the specified field and value. Substitute 'U2' by a valid country subclass. The fields can be anyone given by the Geo::Postcodes::U2::get_fields() call, and the value either a literal text or a regular expression. The resulting list of postcodes is sorted. An empty list is also returned if one ore more of the arguments are invalid, but this can be checked by using the verify_selectionlist procedure described later on.

It is possible to specify more than one field/value pair, but then the mode should be given (but it will default to 'and' otherwise). Use as many field/value-pairs as required. The mode can be specified initially, between the field/value pairs, or not at all.

The following examples are equivalent:

Geo::Postcodes::U2::selection('and', $field => $value, $field2 => $value2);
Geo::Postcodes::U2::selection($field => $value, 'and', $field2 => $value2);
Geo::Postcodes::U2::selection($field => $value, $field2 => $value2);

The field/value pairs are evaluated in the specified order, and the modes can be mixed freely:

Geo::Postcodes::U2::selection($field1 => $value1,
                    'and',    $field2 => $value2,
                    'or',     $field3 => $value3,
                    'or not', $field3 => $value3,
                    'and not, $field4 => $value4,
                    'xor',    $field5 => $value5);
all

All the postcodes. This mode is only legal as the first argument, and any additional arguments are silentliy ignored.

my @postcodes = Geo::Postcodes::U2::selection('all');

This will return all the postcodes.

This is the same as sort get_postcodes(). The object oriented version (see below for syntax) will return a postcode object for each postcode, and can be handy in some circumstances - if time and memory usage is of no concern. Otherwise create the postcode objects only when needed, inside a foreach-loop on the procedure version - or use selection_loop.

and

The postcode is included in the result if it is included in all the expressions.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'and', $field2 => $value2);

Return postcodes matching all the field/value pairs.

The computation will work faster if the field/value pairs are given with the one with the most matches first, and the one with the least matches last. given first

and not

The postcode is included in the result if it is included in the first expression, but not the second one.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'and not', $field2 => $value2);

Return the postcodes not matching any of the field/value pairs. (This is the same as all - or, on sets of postcodes.)

nor

The postcode is included in the result if it is included in none of the expressions.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'nor', $field2 => $value2);
nor not

The postcode is included in the result if it is included in the second expression only.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'nor not', $field2 => $value2);
none

This will return absolutely nothing.

This mode is only legal as the first argument, and any additional arguments are silentliy ignored.

my @postcodes = Geo::Postcodes::U2::selection('none');

This will return undef.

not

This mode can be used initially (as the first argument) to negate the first field/value pair. It is also possible to use 'and not' or any other 'xxx not'-mode initially.

Note that 'not' is not a valid mode, and it will default to 'and' for any additional field/value pairs if no mode is given.

The following examples are equivalent:

Geo::Postcodes::U2::selection('not', $field => $value, 'and', $field2 => $value2);
Geo::Postcodes::U2::selection('not', $field => $value, $field2 => $value2);
Geo::Postcodes::U2::selection('or not', $field => $value, 'and', $field2 => $value2);
Geo::Postcodes::U2::selection('and not', $field => $value, 'and', $field2 => $value2);

The following examples are equivalent:

Geo::Postcodes::U2::selection('or not', $field => $value, $field2 => $value2);
Geo::Postcodes::U2::selection('not', $field => $value, 'or not', $field2 => $value2);
one

This mode can be used initially to limit the returned list of postcodes to just one (or zero). The returned postcode is chosen randomly from the result list.

Geo::Postcodes::U2::selection('one', $field => $value);

It can also be used on its own, just to get a random postcode.

Geo::Postcodes::U2::selection('one');
or

The postcode is included in the result if it is included in at least one of the expressions.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'or', $field2 => $value2);

Return postcodes matching one or more of the field/value pairs.

The computation will work faster if the field/value pairs are given with the one with the least matches first, and the one with the most matches last. given first

or not

The postcode is included in the result if it is included in the first expression, but not the second.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'or not', $field2 => $value2);

It is also possible to achieved this by using 'or' and a reversed regular expression.

xor (exlusive or)

The postcode is included in the result if it is included in only one of the expressions.

my @postcodes = Geo::Postcodes::U2::selection(
   $field1 => $value1, 'xor', $field2 => $value2);
xor not

The postcode is included in the result if it is included in the first but not the second the expressions.

The search value is parsed as the regular expression m{^$value$}i. This has the following implications:

m{...}i

The trailing i means that the search is done case insensitive. This does not work for the special norwegian and danish characters 'Æ', 'Ø' and 'Å' (as used in the subclasses 'NO' and 'DK') unless a use locale is used in the program, and the current locale supports these characters. (This will probably not work when the code is running with mod_perl on a web server.)

'As' will match 'AS', 'As', 'aS', and 'as'.

m{^...$}

The first (^; caret) and last ($; dollar sign) character inside the expression force a matcth to the whole expression.

'AS' will match exactly the four variations mentioned above, and nothing else (so that 'CHAS' or 'ASIMOV' will not match).

Use "Wildcards" or "Regular expressions" to match several characters at once.

Wildcards

The character % (the percent character) will match zero or more arbitrary characters (and is borrowed from standard SQL).

'%12' will match '1112' but not '1201'. 'O%D' will match all strings starting with an 'O' and ending with a 'D'. '%A%' will match all strings with an 'A' somewhere.

(The character % is changed to the regular expression '.*' (dot star) by the module.)

Regular expressions

.

The character . (a single dot) will match exactly one character.

'..11' will match a four character string, where the first two can be anything, followed by '11'.

?

The character ? (a question mark) will match the previous character zero or one time.

'%ØYA?' will match strings ending with 'ØY' or 'ØYA'.

'*'

The character * (a star) will match the previous character zero or more times.

[]

The expression '[AB]' will match one of 'A' or'B'.

%'[AB]' will match all names ending with an 'A' or 'B'.

()

The expression '(..)' will remember the part inside the paranthesis. See the next item for usage.

It can also be used in combination with back references; \1, \2 and so on. (..)\1 will match postcodes starting with two caharcters, and ending with the same ones (e.g. '1919', 7272', but not '1221'). (.)(.)\2\1 will match postcodes where the first and fourth digit is the same, and the second and third digit is the same.

|

The expression 'A|BBB' will match either 'A' or 'BBB'.

Be careful with this construct, as '%ÅS|%SKOG' will not match '%ÅS' or '%SKOG', but rather everything starting with 'ÅS' or ending with 'SKOG' - caused by the '^...$' that the expression is wrapped in. Use '%(ÅS|SKOG)' to get the desired result.

External Procedures

It is also possible to call external procedures, and have them decide if the postcode should be included. Specify 'procedure', \&procedure_to_do_the_job instead of a field/value pair.

The procedure is passed the postcode, and must return true or false.

sub procedure_to_do_the_job
{
  my postcode = shift;
  return 1 if ...
  return 0;
}

selection method

my @postcodobjects = Geo::Postcodes::U2->selection(xxxx);

This works just as the procedure version (see above), but will return a list of postcode objects (instead of just a list of postcodes).

SELECTION_LOOP

It is possible, even adviceable, to use external procedures in the argumnet list to the selection call itself, if the procedure is only used to select from the postcodes. The procedure shown in this section have a print statement, and is okay - if somewhat contrived.

selection_loop procedure

The first argument is a pointer to a procedure which will be called for each postcode returned by the selection call, with the rest of the arguments.

sub post_check
{
  my $postcode = shift;
  print "$postcode\n" if something($postcode) > 3.14;
}

Geo::Postcodes::U2::selection_loop(\&post_check, xxx, yyy);

selection_loop method

As above, but the value passed to the specified procedure is a postcode object.

sub post_check
{
  my $object = shift;
  print $object->postcode() . "\n" if something($object->postcode()) > 3.14;
}

Geo::Postcodes::U2->selection_loop(\&post_check, xxx, yyy);

SUPPORTING PROCEDURES

verify_selectionlist

Use this procedure from the child class to verify that the arguments are valid for use by the selection procedure/method.

my($status, @list) = Geo::Postcodes::U2::verify_selectionlist(@arguments);

A status value of true (1) is followed by a modified version of the original arguments. This will replace things as 'and' 'not' by 'and not', as the selection procedure does not accept the former.

A status value of false (0) is followed by a list of diagnostic messages, up to the point where the verification failed.

External procedures are recognised, and must actually exist for the test to accept them.

Returns true if the mode is one of the list returned by get_selectionmodes, documented below.

Returns true if the mode is one of the list returned by get_initial_selectionmodes, documented below.

Geo::Postcodes::get_selectionmodes

A sorted list of legal selection modes; 'and', 'and not', 'nor', 'nor not', 'or', 'or not', 'xor' and 'xor not'.

Geo::Postcodes::get_initial_selectionmodes

As above, with the addition of 'all', none', 'not' and 'one'. The list is sorted.

SEE ALSO

The tutorial perldoc Geo::Postcodes::Tutorial or man Geo::Postcodes::Tutorial, and the documentation for the individual country modules; e.g. perldoc Geo::Postcodes::NO or man Geo::Postcodes::NO.

AUTHOR

Arne Sommer, <perl@bbop.org>

LICENSE

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

1 POD Error

The following errors were encountered while parsing the POD:

Around line 193:

Non-ASCII character seen before =encoding in ''Æ','. Assuming CP1252