Aggregate pattern: Difference between revisions

Revision as of 13:53, 10 October 2005

Latest revision as of 02:53, 12 February 2025

(44 intermediate revisions by 32 users not shown)

Line 1:

{{Short description|Concepts in statistics and computer science}}

An '''aggregate pattern''' is an important statistical concept in many fields that rely on statistics to predict the behavior of large groups, based on the tendencies of subgroups to consistently behave in a certain way. It is particularly useful in [[sociology]], [[economics]], [[psychology]], and [[criminology]].

{{one source |date=March 2024}}

An '''Aggregate pattern''' can refer to concepts in either statistics or computer programming. Both uses simplify complexity into smaller, simpler parts.

== Statistics ==

In [[computer programming]], an '''aggregate pattern''' is a [[design pattern (computer science)|design pattern]].

An aggregate pattern is an important statistical concept in many fields that rely on [[statistics]] to predict the behavior of large groups, based on the tendencies of subgroups to consistently behave in a certain way. It is particularly useful in [[sociology]], [[economics]], [[psychology]], and [[criminology]].

== Computer programming ==

Members of a common subclass are each known to have certain [[method (computer science)|method]]s. These methods return information about the state of that particular object. It does happen that an application is concerned with an aggregation, or amalgamation, of data from several object of the same type. This leads to code being repeated around the program:

In ''[[Design Patterns]]'', an aggregate is not a [[Software design pattern|design pattern]] but rather refers to an object such as a list, vector, or generator which provides an interface for creating [[iterator]]s. The following example code is in [[Python (programming language)|Python]].

my $subtotal;

<syntaxhighlight lang="python">

foreach my $item (@cart) {

def fibonacci(n: int):

$subtotal += $item->query_price();

a, b = 0, 1

}

count = 0

while count < n:

my $weight;

count += 1

foreach my $item (@cart) {

a, b = b, a + b

$weight += $item->query_weight();

yield a

}

# and so on

for x in fibonacci(10):

== Representing ==

print(x)

Representing individual objects when the application is concerned about the general state of several objects is an ImpedenceMismatch. This is a common mismatch as programmers feel obligated to model the world in minute detail then are pressed with the problem of giving it all a high level interface.

def fibsum(n: int) -> int:

Create an object as a [[wrapper]], using the same API, with a common subtype

total = 0

as a cart entry, but allow it to hold other objects of that subtype: make

for x in fibonacci(n):

it a container. Define its accessors to return aggregate information

total += x

on the objects it contains.

return total

def fibsum_alt(n: int) -> int:

package Cart::Basket;

"""

Alternate implementation. demonstration that Python's built-in function sum()

@ISA = qw(Cart::Item);

works with arbitrary iterators.

"""

sub query_price {

return sum(fibonacci(n))

my $self = shift;

my $contents = $self->{contents};

foreach my $item (@$contents) {

}

}

# other query_ routines here...

sub add_item {

my $self = shift;

my $contents = $self->{contents};

my $item = shift; $item->isa('Cart::Item') or die;

push @$contents, $item;

return 1;

}

myNumbers = [1, 7, 4, 3, 22]

The aggregation logic, in this case, totalling, need only exist in this

container, rather than being strewn around the entire program. Less code,

less [[code momentum]], fewer dependencies, more flexibility.

def average(g) -> float:

We have an object of base type //Cart::Item// that itself holds other //Cart::Item//

return float(sum(g)) / len(g) # In Python 3 the cast to float is no longer be necessary

objects. That makes us recursive and nestable - one basket could hold several

</syntaxhighlight>

items along with another basket, into which other items and baskets could

Python hides essentially all of the details using the [https://docs.python.org/3/library/stdtypes.html#iterator-types iterator protocol]. Confusingly, ''[[Design Patterns]]'' uses "aggregate" to refer to the blank in the code <code>for x in ___:</code> which is unrelated to the term "aggregation".<ref>[[Design Patterns]], p. 22: "Aggregation implies that one object owns or is responsible for another object. ... Aggregation implies that an aggregate object and its owner have identical lifetimes."</ref> Neither of these terms refer to the statistical aggregation of data such as the act of adding up the Fibonacci sequence or taking the average of a list of numbers.

be placed. You may or may not want to do this intentionally, but to someone

casually calling //->query_price()// on your //Cart::Basket// object

won't have to concern himself with this - things will just work.

== See also ==

This will break. Unless the advice of AbstractRootClasses is followed and

different implementations of the same thing share the same interface, the

basket can't confidently aggregate things. Unless the advice of StateVsClass

is heeded, [[abstract root classes]] will never be achieved: the temptation to

draw distinctions between classes that lack certain functions will be too

strong. These distinctions run counter to AbstractRootClasses, causing

segmentation and proliferation of interfaces for no good reason. This proliferation

of types prevents aggregation in baskets and containers. Avoid this vicious

cycle. Parrots that don't squak are still parrots.

XXX IteratorInterface blurb - aggregation is kind of like iteration in

that they both present information gleaned from a number of objects through

a tidy interface in one object. While IteratorInterface deals with each

contained or known object in turn, AggregatePattern summarizes them in

one fell swoop.

''The article is originally from [[Perl Design Patterns Book]].''

==See also==

* [[Visitor pattern]]

* [[Template class]]

* [[Facade pattern]]

* [[Iterator interface]]

* [[Type safety]]

* [[Functional programming]]

* [[State vs class]]

== References ==

[[Category:Programming]]

{{Reflist}}

[[Category:Software design patterns]]

[[Category:Articles with example Python (programming language) code]]

{{statistics-stub}}

{{compu-prog-stub}}