Sorting(3pm) | User Contributed Perl Documentation | Sorting(3pm) |
Data::Sorting - Multi-key sort using function results
use Data::Sorting qw( :basics :arrays :extras ); # Sorting functions default to simple string comparisons @names = qw( Bob Alice Ellen Charlie David ); @ordered = sorted_by( undef, @names ); # Various options can be passed before the list values @ordered = sorted_by( [ -order=>'reverse' ], @names ); # You can also generate a sorting function and then apply it $function = sort_function(); @ordered = $function->( @names ); # or &{$function}(@names) @ordered = sort_function( -order=>'reverse' )->( @names ); # The :array functions are prototyped to take the array first @ordered = sorted_array( @names ); @ordered = sorted_arrayref( \@names ); # You can also sort an array in place, changing its internal order sort_array( @names ); sort_arrayref( \@names ); # There are several sorting options, such as -compare => 'natural' @movies = ( 'The Matrix', 'Plan 9', '2001', 'Terminator 2' ); @ordered = sort_function( -compare => 'natural' )->( @movies ); # @ ordered now contains '2001', 'The Matrix', 'Plan 9', 'Terminator 2' # To sort numbers, pass the -compare => 'numeric' option @numbers = ( 18, 5, 23, 42, 156, 91, 64 ); @ordered = sorted_by( [ -compare => 'numeric' ], @numbers ); @ordered = sort_function( -compare => 'numeric' )->( @numbers ); @ordered = sorted_array( @numbers, -compare => 'numeric' ); sort_array( @numbers, -compare => 'numeric' ); # You can sort by the results of a function to be called on each item sort_array( @numbers, -compare => 'numeric', sub { $_[0] % 16 } ); # @numbers now contains 64, 18, 5, 23, 42, 91, 156 # For arrays of datastructures, pass in keys to extract for sorting @records = ( { 'rec_id'=>3, 'name'=>{'first'=>'Bob', 'last'=>'Macy'} }, { 'rec_id'=>1, 'name'=>{'first'=>'Sue', 'last'=>'Jones'} }, { 'rec_id'=>2, 'name'=>{'first'=>'Al', 'last'=>'Jones' } }, ); @ordered = sorted_array( @records, 'rec_id' ); # For nested data structures, pass an array of keys to fetch @ordered = sorted_array( @records, ['name','first'] ); # Pass multiple sort keys for multiple-level sorts @ordered = sorted_array( @records, ['name','last'], ['name','first'] ); # Any selected sort options are applied to all subsequent sort keys @ordered = sorted_array( @records, -order => 'reverse', ['name','last'], ['name','first'] ); # Options specified within a hash-ref apply only to that key @ordered = sorted_array( @records, { order=>'reverse', sortkey=>['name','last'] }, ['name','first'] ); # Locale support is available if you have Perl 5.004 or later and POSIX POSIX::setlocale( POSIX::LC_COLLATE(), 'en_US' ); POSIX::setlocale( POSIX::LC_CTYPE(), 'en_US' ); @ordered = sorted_array( @records, -compare=>'locale', ['name','last'], ['name','first'] );
Data::Sorting provides functions to sort the contents of arrays based on a collection of extraction and comparison rules. Extraction rules are used to identify the attributes of array elements on which the ordering is based; comparison rules specify how those values should be ordered.
Index strings may be used to retrieve values from array elements, or function references may be passed in to call on each element. Comparison rules are provided for numeric, bytewise, and case-insensitive orders, as well as a 'natural' comparison that places numbers first, in numeric order, followed by the remaining items in case-insensitive textual order.
This module provides several public functions with different calling interfaces that all use the same underlying sorting mechanisms.
These functions may be imported individually or in groups using the following tags:
All of these functions take a list of sorting rules as arguments. See "Sort Rule Syntax" for a discussion of the contents of the $sort_rule or @sort_rules parameters shown below.
@ordered = sorted_by( $sort_rule, @value_array ); @ordered = sorted_by( $sort_rule, @$value_arrayref ); @ordered = sorted_by( $sort_rule, $value1, $value2, $value3 ); @ordered = sorted_by( \@sort_rules, @value_array ); @ordered = sorted_by( \@sort_rules, @$value_arrayref ); @ordered = sorted_by( \@sort_rules, $value1, $value2, $value3 );
This is a general-purpose sorting function which accepts one or more sort order rules and a list of input values, then returns the values in the order specified by the rules.
@ordered = sort_function( @sort_rules )->( @value_array ); @ordered = sort_function( @sort_rules )->( @$value_arrayref ); @ordered = sort_function( @sort_rules )->( $value1, $value2, $value3 );
Creates an anonymous function which applies the provided sort rules. The function may be cached and used multiple times to apply the same rules again.
@ordered = sorted_array( @value_array, @sort_rules ); @ordered = sorted_array( @$value_arrayref, @sort_rules );
Returns a sorted list of the items without altering the order of the original list.
@ordered = sorted_arrayref( \@value_array, @sort_rules ); @ordered = sorted_arrayref( $value_arrayref, @sort_rules );
Returns a sorted list of the items without altering the order of the original list.
sort_array( @value_array, @sort_rules ); sort_array( @$value_arrayref, @sort_rules );
Sorts the contents of the specified array using a list of sorting rules.
sort_arrayref( \@value_array, @sort_rules ); sort_arrayref( $value_arrayref, @sort_rules );
Equivalent to sort_array, but takes an explicit array reference as its first argument, rather than an array variable.
@key_values = sort_key_values( \@value_array, @sort_rules ); @key_values = sort_key_values( $value_arrayref, @sort_rules );
Doesn't actually perform any sorting. Extracts and returns the values which would be used as sort keys from each item in the array, in their original order.
@description = sort_description( $descriptor, @sort_rules );
Doesn't actually perform any sorting. Provides descriptive information about the sort rules for diagnostic purposes.
The sort rule argument list may contain several different types of parameters, which are parsed identically by all of the public functions described above.
A sort rule definition list may contain any combination of the following argument structures:
@ordered = sorted_array( @names );
@ordered = sorted_array( @numbers, sub { $_[0] % 8 } ); @ordered = sorted_array( @records, 'rec_id' ); @ordered = sorted_array( @records, ['name','first'] );
Any number of sortkeys may be provided:
@ordered = sorted_array( @records, ['name','last'], ['name','first'] );
@ordered = sorted_array( @numbers, -sortkey => sub { $_[0] % 8 } ); @ordered = sorted_array( @records, -sortkey => ['name','last'], -sortkey => ['name','first'] );
@ordered = sorted_array( @numbers, { sortkey => sub { abs(shift) }, compare => 'numeric', } );
@ordered = sorted_array( @records, -compare => 'numeric', -sortkey => sub { abs(shift) }); @ordered = sorted_array( @records, -compare => 'textual', -sortkey => ['name','last'], -sortkey => ['name','first'] );
The possible option values are:
Each of these options is discussed at further length below.
For the extract option, you may specify one of the following option_values:
If the sortkey is an array reference, then the keys are looked up sequentially, allowing you to sort on the contents of a nested hash or array structure.
If the sortkey is an array reference, then the first value is used as the method name and the remaining values as arguments to that method.
If the sortkey is an array reference, then the first value is used as the function reference and the remaining values as arguments to be passed after the item value.
extract => self | method | key | code | CODEREF | ... sortkey => - | m.name | key/idx | CODEREF | args
For the compare option, you may specify one of the following option_values:
Each of these functions may return a postive, zero, or negative value based on the relationship of the values in the $a and $b positions of the current @ValueSet array. An undefined return indicates that the comparator is unable to provide an ordering for this pair, in which case the choice will fall through to the next comparator in the list; if no comparator specifies an order, they are left in their original order.
For the order option, you may specify one of the following option_values:
Depending on the specific sorting rules used in a given call, this module automatically selects an internal function that provides an appropriate approach to implementing the sort, called the sort "engine".
You can override this selection by setting an "engine" option on the first sort key, which can either contain either the name of one of the engines, described below, or a CODEREF with equivalent behavior.
You may also set the $PreCalculate package variable to true to force this engine to be selected. Because the sort key values for the list are calculated before entering Perl's sort operation, there's less of a chance of possible re-entry problems due to nested uses of the sort operator, which causes a fatal error in at least some versions of Perl.
This release of Data::Sorting is intended for public review and feedback.
Name DSLIP Description -------------- ----- --------------------------------------------- Data:: ::Sorting bdpfp Multi-key sort using function results
Further information and support for this module is available at www.evoscript.org.
Please report bugs or other problems to <bugs@evoscript.com>.
The following issues have been noted for future improvements:
Convert more types of comparisons to packed-default sorts for speed.
Further investigate the current status of the Sort::Records module.
Add a comparator function for an alpha-numeric-spans sorting model like Sort::Naturally.
Interface to Sort::PolySort for alternate comparator styles, like "name" and "usdate".
For non-scalar values, compare referents along the lines of Ref::cmpref().
Provide better handling for nested sorts; perhaps throw an exception from the inner instance to the outer, catch and set $PreCalculate, then go back into the loop?
Replace dynamic scoping with object instances for thread safety. May not be necessary given changes in threading models.
M. Simon Cavalletto, simonm@cavalletto.org Evolution Softworks, www.evoscript.org
Copyright 2003 Matthew Cavalletto.
Portions copyright 1996, 1997, 1998, 1999 Evolution Online Systems, Inc.
You may use, modify, and distribute this software under the same terms as Perl.
2022-11-21 | perl v5.36.0 |