迭代器（Iterator）模式，又叫做游标（Cursor）模式。提供一种方法访问一个容器（container）或者聚集（Aggregator）对象中各个元素，而又不需暴露该对象的内部细节。在采用不同的方式迭代时，只需要替换相应Iterator类即可。本文采用Matlab语言实现对元胞数组和string数组的遍历。
Aggregator.m
classdef Aggregator < handle
methods(Abstract)
iterObj = createIterator(~);
end
end
CellAggregator.m
classdef CellAggregator < Aggregator
properties
cell
end
methods
function obj = CellAggregator(cell)
obj.cell = cell;
end
function iterObj = createIterator(obj)
iterObj = CellIterator(obj);
end
end
end
StringAggregator.m
classdef StringAggregator < Aggregator
properties
string_arr
end
methods
function obj = StringAggregator(string_arr)
obj.string_arr = string_arr;
end
function iterObj = createIterator(obj)
iterObj = StringIterator(obj);
end
end
end
Iterator.m
classdef Iterator < handle
methods(Abstract)
hasNext(~);
next(~);
end
end
CellIterator.m
classdef CellIterator < Iterator
properties
index = 1;
aggHandle;
end
methods
function obj = CellIterator(agg)
obj.aggHandle = agg;
end
function res = hasNext(obj)
if(obj.index <= length(obj.aggHandle.cell))
res = true;
else
res = false;
end
end
function ele = next(obj)
if(obj.hasNext())
ele = obj.aggHandle.cell{obj.index};
obj.index = obj.index + 1;
else
ele = [];
end
end
end
end
StringIterator.m
classdef StringIterator < Iterator
properties
index = 1;
aggHandle;
end
methods
function obj = StringIterator(agg)
obj.aggHandle = agg;
end
function res = hasNext(obj)
if(obj.index <= obj.aggHandle.string_arr.length)
res = true;
else
res = false;
end
end
function ele = next(obj)
if(obj.hasNext())
ele = obj.aggHandle.string_arr(obj.index);
obj.index = obj.index + 1;
else
ele = string.empty();
end
end
end
end
测试代码：
cell = CellAggregator({'matlab','cell','iter'});
iterObj = cell.createIterator();
while iterObj.hasNext()
disp(iterObj.next());
end
str_arr = StringAggregator(["matlab","string","iter"]);
iterObj = str_arr.createIterator();
while iterObj.hasNext()
disp(iterObj.next());
end
迭代器模式不仅可以对集合进行顺序遍历，而且可以以用户想要的条件进行遍历，如果想要统计列表中所有偶数的和，可以写成循环语句，如下所示：
test_for.m
list = [13,8,17,11,8,19,18,12,13,12]; sum = 0;
for i=1:length(list)
if mod(list(i),2) == 0
sum = sum + list(i);
end
end disp("列表中所有偶数的和:" + sum); sum = 0;
for i=1:length(list)
if mod(list(i),2) == 1
sum = sum + list(i);
end
end disp("列表中所有奇数的和:" + sum);
如果要统计列表中所有奇数的和，这就需要对循环语句块进行修改，如上所示。这就违反了对修改关闭的原则。而迭代器模式可以将循环语句块和业务语句块（sum求和）分离出去，这就需要对Aggregator和Iterator类进行继承，如下所示：
ListAggregator.m
classdef ListAggregator < Aggregator
properties(Constant)
EVEN_ITER = 0
ODD_ITER = 1
end
properties
list
end
methods
function obj = ListAggregator(list)
obj.list = list;
end
function iterObj = createIterator(obj,type_iter)
if(type_iter == ListAggregator.EVEN_ITER)
iterObj = EvenIterator(obj);
else
iterObj = OddIterator(obj);
end
end
end
end
OddIterator.m
classdef OddIterator < Iterator
properties
index = 1;
aggHandle;
end
methods
function obj = OddIterator(agg)
obj.aggHandle = agg;
end
function res = hasNext(obj)
res = false;
while(obj.index <= length(obj.aggHandle.list))
if mod(obj.aggHandle.list(obj.index),2) == 1
res = true;
break;
end
obj.index = obj.index + 1;
end
end
function ele = next(obj)
if(obj.hasNext())
ele = obj.aggHandle.list(obj.index);
obj.index = obj.index + 1;
else
ele = [];
end
end
end
end
EvenIterator.m
classdef EvenIterator < Iterator
properties
index = 1;
aggHandle;
end
methods
function obj = EvenIterator(agg)
obj.aggHandle = agg;
end
function res = hasNext(obj)
res = false;
while(obj.index <= length(obj.aggHandle.list))
if mod(obj.aggHandle.list(obj.index),2) == 0
res = true;
break;
end
obj.index = obj.index + 1;
end
end
function ele = next(obj)
if(obj.hasNext())
ele = obj.aggHandle.list(obj.index);
obj.index = obj.index + 1;
else
ele = [];
end
end
end
end
test_iter.m
list_agg = ListAggregator([13,8,17,11,8,19,18,12,13,12]);
iterObj = list_agg.createIterator(ListAggregator.EVEN_ITER);
sum = 0;
while iterObj.hasNext()
sum = sum + iterObj.next();
end
disp("列表中所有偶数的和:" + sum); iterObj = list_agg.createIterator(ListAggregator.ODD_ITER);
sum = 0;
while iterObj.hasNext()
sum = sum + iterObj.next();
end
disp("列表中所有奇数的和:" + sum);
如果需要返回列表标，可以将next函数的返回值定义成[ele,index] = next(obj)，python里面有yield关键字，可以让迭代器模式更加简洁，见参考资料。
迭代器模式还可以扁平化多层嵌套循环，如果找到所有水仙花数，可以使用三层循环，代码如下：
test_for.m
a = 1:9;
b = 0:9;
c = 0:9; for i=1:length(a)
for j = 1:length(b)
for k = 1:length(c)
num = a(i)*100 + b(j)*10 + c(k);
if a(i)^3 + b(j)^3 + c(k)^3 == a(i)*100 + b(j)*10 + c(k)
disp("" + num + "是水仙花数");
end
end
end
end
如果使用迭代器模式来简化三层循环，只需要对Aggregator和Iterator类进行继承，代码如下：
MutiListAggregator.m
classdef MutiListAggregator < Aggregator
properties
lists
end
methods
function obj = MutiListAggregator(varargin)
obj.lists = varargin;
end
function iterObj = createIterator(obj)
iterObj = MutiListIterator(obj);
end
end
end
MutiListIterator.m
classdef MutiListIterator < Iterator
properties
indexs = [];
lengths = [];
aggHandle;
end
methods
function obj = MutiListIterator(agg)
obj.aggHandle = agg;
for i= 1:length(obj.aggHandle.lists)
obj.lengths(i) = length(obj.aggHandle.lists{i});
end
obj.indexs = ones(1,length(obj.aggHandle.lists));
end
function res = hasNext(obj)
if(obj.indexs(1) > obj.lengths(1))
res = false;
else
res = true;
end
end
function varargout = next(obj)
if(obj.hasNext())
for i = 1:length(obj.aggHandle.lists)
temp_list = obj.aggHandle.lists{i};
varargout{i} = temp_list(obj.indexs(i));
end
obj.indexs(length(obj.aggHandle.lists)) = obj.indexs(length(obj.aggHandle.lists)) + 1;
for i = length(obj.aggHandle.lists):-1:2
if(obj.indexs(i) > obj.lengths(i))
obj.indexs(i) = mod(obj.indexs(i), obj.lengths(i));
obj.indexs(i - 1) = obj.indexs(i - 1) + 1;
end
end
else
for i = 1:length(obj.aggHandle.lists)
varargout{i} = [];
end
end
end
end
end
test_iter.m
list_agg = MutiListAggregator(1:9,0:9,0:9);
iterObj = list_agg.createIterator();
while iterObj.hasNext()
[a,b,c] = iterObj.next();
num = a * 100 + b * 10 + c;
if a^3 + b^3 + c^3 == a * 100 + b * 10 + c
disp("" + num + "是水仙花数");
end
end
参考资料：
https://mp.weixin.qq.com/s/tghaVkTPUijBJO0qxSXdZQ