NAME Set::ArrayAlt - full set of list/set operations for arrays SYNOPSIS $obj = new Set::ArrayAlt [@list]; DESCRIPTION *** NOTE *** This module has been renamed to Array::AsObject . Please use that module instead. For more information, please refer to the HISTORY AND RATIONALE SECTION below. *** END NOTE *** This module attempts to provide a complete set of set and list operations. It acts as a standalone module, or is available in Template::Toolkit templates using the companion module Template::Plugin::ListOps . It performs many of the same functions as the Set::Array module, but has better support for handling duplicate elements in lists, and does not require the Want module. For more details, please refer to the HISTORY AND RATIONALE section below. BASE OBJECT METHODS new $obj = new Set::ArrayAlt [@list]; This creates a new Set::ArrayAlt object. If @list is passed in, this is set to the initial set of elements. If @list is passed in, the following two are equivalent: $obj = new Set::ArrayAlt @list; and $obj = new Set::ArrayAlt; $obj->list(@list); There is no restriction on the type of elements in @list. They can be scalars, references, objects, or undefs. version $vers = $obj->version(); Gets the version of this module. err errmsg $err = $obj->err(); Check to see if the previous operation produced an error. $msg = $obj->errmsg(); If an operation did produce an error, this will get the error message. LIST EXAMINATION METHODS The following methods examine a list but do not modify the object. They will all produce an error if the list has not been defined and undef is returned. as_hash %hash = $obj->as_hash(); This returns a hash that describes the scalars in the list contained in $obj. References and undef values are ignored. Every scalar in the list is one of the keys in %hash. The value of %hash is the number of times that element appears in the list. ($count,$vals) = $obj->as_hash(1); This returns two hash references that describe ALL values in the list, including references, objects, and undef. Every unique value in the list is assigned a label (which has no significance). The hash %$vals has the labels as keys and the actual values as values for the keys. The hash %$count has the labels as keys and the values are the number of times that value appears in the list. The following example should illustrate this: $i = [1,2]; $j = [1,2]; @l = ('a', $i, 'b', $j, 'a', undef, undef, $i); $o = new Set::ArrayAlt(@l); ($c,$v) = $o->as_hash(1) => $c = { 1 => 2, 2 => 2, 3 => 1, 4 => 1, 5 => 2 } $v = { 1 => 'a', 2 => [1,2], 3 => 'b', 4 => [1,2], 5 => undef } Note that the two elements $i and $j are treated as different since they point to different references, even though the data is the same in those two lists. at $ele = $obj->at($n); @ele = $obj->at(@n); This returns the element at position $n in the list, or a list of the elements at the positions given in the list @n. In scalar context, only a single position may be passed in. In list context, any number of positions may be passed in. Positions follow standard perl conventions with numbering starting at 0. Negative numbers can also be used to count from the end. All positions must refer to elements in the list (i.e. you may not refer to the 6th element in a list containing 5 elements). count $num = $obj->count([$val]); This counts the number of times val appears in the list. If $val is not given, it counts the number of undef elements in the list. exists $flag = $obj->exists(@val); This returns 1 if every value exists in the list. If @val is not passed in, it tests for undef values. It returns 0 if any of the values do not exists in the list. first last $val = $obj->first(); $val = $obj->last(); These return the first or last elements of the list. If the list contains no elements, an error is set. index rindex $idx = $obj->index([$val]); @idx = $obj->index([$val]); $idx = $obj->rindex([$val]); @idx = $obj->rindex([$val]); In list context, these return the index of every occurrence of $val in the list. If $val is not passed in, the indices of all undef elements are returned. If the value is not found, -1 is returned in scalar context, or an empty list in list context. The rindex function returns them in reverse order. In scalar context, the index and rindex methods return the index of the first or last occurrence of $val in the list. $val can be a scalar, undef, or a reference, and all will work as expected. For example: $i = [1,2]; $j = [1,2]; @l = ('a', $i, 'b', $j, 'a', undef, undef, $i); $o = new Set::ArrayAlt(@l); (@idx) = $o->index(); => @idx = (5, 6) (@idx) = $o->index($i); => @idx = (1) is_empty $flag = $obj->is_empty([$undef]); This checks to see if the list is empty. If $undef is not passed in, a list is empty only if the length is 0. If $undef is passed in, the list is also empty if it consists only of undef values. length $num = $obj->length(); Returns the number of elements in the list. list @list = $obj->list(); This returns the list stored in the object. SIMPLE LIST MODIFICATION METHODS The following methods will modify the object. They will all produce an error if the list has not been defined and undef is returned. clear $obj->clear(); This removes all elements from the list (sets it to a zero-length list). $obj->clear(1); This sets all elements in the list to be undef (preserving the length). The list must be defined or an error results. compact $obj->compact(); This removes any undef objects from the list. delete $obj->delete($all,$undef,@val); This deletes occurences of each values from the list. If $all is 1, all occurences are removed. Otherwise, only the first occurence of each value is removed. If $undef is 1, values are replaced with undef. Otherwise, they are completely removed. delete_at $obj->delete($undef,@idx); This deletes elements at the given indices. The order of the indices is not important. They will be deleted in the order of last to first. If $undef is 1, values are replaced with undef. Otherwise, they are completely removed. fill $obj->fill([$val,] [$start,] [$length]); This sets elements of a list to be $val. If $val is not passed in, values are set to undef. $start can be a positive or negative number. It must be an index in the list. It can also be the index of the first element after the list. So, if the list contains 3 elements, $start can be -3 to +3. Negative values refer to the index from the end as usual. 0 to 2 refer to the index of the elements in the list, and 3 is the first element after the list. $start defaults to 0. $length can be any positive value and refers to the number of elements that will be set to the value. If $length is omitted, it defaults to the number of elements in the list starting at $start, or to 1 if $start is the first element after the list. So if list contains 3 elements, and $start is 1, $length will default to 2 (the number of elements in the list starting at index 1. list $obj->list(@list); This sets the object to contain the given list. Any previous list is replaced. min max $ele = $obj->min([$method [,@args]]); $ele = $obj->max([$method [,@args]]); These return the first/last value from the list if it were sorted with the given method using the Sort::DataTypes module. By default, if $method is not given, the numerical minimum/maximum value is given. Otherwise, $method can be any sort method available from the Sort::DataTypes module. For example, to get the first word alphabetically, use $ele = $obj->min("alphabetic"); pop shift $val = $obj->pop(); $val = $obj->shift(); These perform the standard pop/shift operations. push unshift $obj->push(@list); $obj->unshift(@list); These perform the standard push/unshift operations. randomize $obj->randomize(); This randomly shuffles the list. reverse $obj->reverse(); This reverses the list. rotate $obj->rotate([$num]); This rotates the list. If $num is not included, it defaults to 1. If $num is a positive number, the first element from the list is removed and pushed on to the end a total of $num times. If $num is a negative number, the last element from the list is removed and shifted onto the front a total of $num times. set $obj->set($index [,$val]); This sets the list index to the given value, or undef if no value is passedin. sort $obj->sort([$method [,@args]]); This uses any method from the Sort::DataTypes module to sort the list. Method can be of the form "numerical" or "rev_numerical" to do forward and reverse sorting. @args may be passed in if the method requires additional arguments. If no method is given, it defaults to alphabetical sorting. splice @vals = $obj->splice([$start,] [$length,] [@list]); This performs the perl splice command on a list. If $start is omitted (or is undefined), it defaults to 0. If $length is omitted (or undefined), it defaults to the end of the list. The values removed are returned, and are replaced with @list if present. unique $obj->unique(); This removes any duplicates in a list. The first occurrence of each element is kept, and the order of those first occurrences is preserved. SET METHODS The following methods work with two Set::ArrayAlt objects. They apply a set operation to the two of them and produce a value or a third Set::ArrayAlt object containing the results. If an error occurs, it is set in the returned object, NOT in any of the original objects. The original objects are unmodified in all cases. difference $obj3 = $obj->difference($obj2 [,$unique]); This takes two lists and removes the second list from the first. By default, one occurence of every element in the second list is removed from the first list. If $unique is given, every element in the first list is removed. For example, the difference of the two lists (a a b b c) and (a) is either (a b b c) or (b b c). If $unique is non-zero, the second is given. It should be noted that both "b" elements in the example will be kept regardless of the value of $unique because the $unique flag only affects elements being removed. intersection $obj3 = $obj->intersection($obj2 [,$unique]); This takes two lists and finds the intersection of the two. The intersection are elements that are in both lists. The returned list is in the order they appear in the first list. By default, duplicate elements are treated individually unless $unique is passed in. For example, the intersection of two lists (a a b b c) and (a a c d) is either (a a c) or (a c). If $unique is non-zero, the second is given. is_equal not_equal $flag = $obj->is_equal($obj2 [,$unique]); $flag = $obj->not_equal($obj2 [,$unique]); These take two lists and test to see if they are equal. If an error is encountered, undef is returned, but no error is stored. The order of the elements is ignored so (a,b) = (b,a). If $unique is non-zero, the count of each type of element is ignored so (a,a,b) = (a,b). Otherwise, the count is important so (a,a,b) != (a,b). is_subset not_subset $flag = $obj->is_subset($obj2 [,$unique]); $flag = $obj->not_subset($obj2 [,$unique]); These return 1 if the list in $obj2 is a subset of the list in $obj (or is NOT a subset). If $unique is not passed in, every element in $obj2 must have an instance in $obj. So (a a b) is a subset of (a a a b c) but NOT of (a b c). If $unique is passed in, every element in $obj2 must exists in $obj but the count is unimportant, so (a a b) is a subset of (a b c). symmetric_difference $obj3 = $obj->symmetric_difference($obj2 [,$unique]); This takes two lists and finds the symmetric difference of the two. The symmetric difference are elements that are in either list, but not both. The order of the list produced are the elements from the first object (order preserved) followed by those from the second object. If $unique is non-zero, one instance of an element cancels out all of the instances in the other list. For example, the symmetric difference between the two lists (a a b b c) and (a c) is either (a b b) or (b b). If $unique is non-zero, the second is used. Note that both instances of 'b' are kept because the $unique flag only affects elements which exist in both objects. union $obj3 = $obj->union($obj2 [,$unique]); This takes two lists and combines them. By default, every element is preserved. If $unique is passed in, the duplicates are removed. For example, the union of the two lists (a a b) and (a c) is either (a a b a c) or (a b c). The second is returned if $unique is non-zero. HISTORY AND RATIONALE With several other modules out there that do various set and array operations, a brief history of why I wrote this module is in order. The origin of this module came when I needed better list handling operations (especially involving lists that might contain duplicate elements) inside of a Template::Toolkit template. The built in list functions in Template::Toolkit weren't sufficent for my needs, so I looked around. The module that came closest to my needs was Set::Array. Although not a perfect match for what I wanted (I really wanted better support for lists with duplicate elements), it came close enough, so I wrote a wrapper module (Template::Plugin::ListOps) around it to do most of what I wanted. Unfortunately, I discovered almost immediately that Set::Array suffered from a fairly serious problem. It depends on the Want module which, at the time, had some known problems and would fail under some circumstances (older versions of perl if I recall correctly, though I could be wrong about that), and unfortunately, some of the places I needed my module to run failed due to those problems. I looked at the Want module, but correcting it was beyond my abiltity, so the best solution seemed to be to rewrite the module without depending on Set::Array. This would also allow me to add the functionality that I wanted. So I did that. I rewrote each function to do the list/set operation I wanted instead of calling Set::Array functions. Almost as soon as I was done (and perhaps even before), I started regretting rewriting the module in that way. I should have written a standalone module and then had the Template::Plugin::ListOps be a wrapper for it instead of Set::Array... but in the interest of time, I didn't go back and redo it... until later. Later, I ran into a case where I wanted the set/list functionality from Template::Plugin::ListOps for another perl application. At that point, I decided to create the standalone module. So, this module takes the routines from the original Template::Plugin::ListOps module and moves them into a standalone module. Template::Plugin::ListOps was then rewritten trivially to be a wrapper around this module. Some other notes: Since the original version of Template::Plugin::ListOps (which was never released) was a wrapper around Set::Array, the naming of the functions is very similar, but the functionality differs slightly. This module was initially named Set::ArrayAlt to indicate that it is based on Set::Array, but with a few changes. It has enhanced functionality with respect to duplicate elements but is missing some of the functionality of Set::Array (especially method chaining and operator overloading) which depend on the Want module. This module is not intended to be a drop-in replacement for Set::Array. It is also missing a couple functions (join and impose) that are applicable only to all-scalar lists. It may well be that the problems with the Want module have been corrected at this point, and that I could have used Set::Array, but since some of the functionality I needed was the enhanced duplicate element handling, and since creating this module from what I'd already written was actually a pretty easy task, I decided to go ahead with the creation of this module. Anyway, that's the history. Hopefully, I'm justified in reinventing the wheel. After a while, I decided I wanted to register this module (which basically means to get it put in the official list of perl modules). The upside is that the module will get added publicity and use... the downside (if you can call it that) is that they expect the name of the module to accurately reflect the module. This module is more accurately thought of as a module for handling arrays, than as a module for handling sets. True, it does do set operations, but it really works on an array, doing most of the operations that you could want to do with an array. Included in those operations are a subset of functions where you treat the array as a set. Anyway, it was requested that I rename the module to be in the Array:: namespace rather than the Set:: namespace (and Array::AsObject was suggested), so that's why it's been renamed. Versions of this module before 1.02 were released under the name Set::ArrayAlt. Version 1.02 was released simultaneously under the names Set::ArrayAlt and Array::AsObject. Versions after 1.02 will only be released under the name Array::AsObject. KNOWN PROBLEMS None at this point. LICENSE This script is free software; you can redistribute it and/or modify it under the same terms as Perl itself. AUTHOR Sullivan Beck (sbeck@cpan.org)