基本介紹
- 迭代器模式(Iterator Pattern)是常用的設(shè)計(jì)模式，屬于行為型模式
- 如果我們的集合元素是用不同的方式實(shí)現(xiàn)的，有數(shù)組，還有java的集合類,或者還有其他方式，當(dāng)客戶端要遍歷這些集合元素的時(shí)候就要使用多種遍歷方式，而且還會(huì)暴露元素的內(nèi)部結(jié)構(gòu)，可以考慮使用迭代器模式解決。.
- 迭代器模式，提供一種遍歷集合元素的統(tǒng)一接口，用一致的方法遍歷集合元素，不需要知道集合對(duì)象的底層表示，即:不暴露其內(nèi)部的結(jié)構(gòu)。
image.png
- Iterator：迭代器接口，系統(tǒng)提供。
- ConcereteIterator：具體的迭代器類，管理迭代。
- Aggregate：一個(gè)統(tǒng)一的接口，將客戶端和具體聚合解耦。
- ConcereteAggregate：具體的聚合，持有對(duì)象集合，并提供一個(gè)方法，返回一個(gè)迭代器。
例子：
需求：一個(gè)學(xué)校有很多學(xué)院，一個(gè)學(xué)院有很多系，展示院系組成。
實(shí)現(xiàn)：
image.png
package iterator;
public class Department {
private String name;
public Department(String name) {
super();
this.name = name;
}
/**
* @return the name
*/
public String getName() {
return name;
}
}
public abstract class College {
private String name;
public College(String name) {
super();
this.name = name;
}
public abstract Iterator<?> createIterator();
public String getName() {
return name;
}
}
public class ComputerCollegeIterator<E> implements Iterator<Object>{
//這里需要知道基本元素是以怎么樣的方式存放的
Department[] departments;
int position = 0;
int length;
public ComputerCollegeIterator(Department[] departments,int length) {
super();
this.departments = departments;
this.length = length;
}
@Override
public boolean hasNext() {
// TODO Auto-generated method stub
if (position >= length) {
return false;
}
return true;
}
@Override
public Department next() {
// TODO Auto-generated method stub
return departments[this.position++];
}
@Override
public void remove() {
// TODO Auto-generated method stub
}
}
public class ComputerCollege extends College{
Department[] departments;
int numOfDepartments = 0;
public ComputerCollege(String name) {
super(name);
// TODO Auto-generated constructor stub
this.departments = new Department[5];
addDepartment("計(jì)算機(jī)");
addDepartment("軟件");
}
public void addDepartment(String name) {
if (numOfDepartments == departments.length) {
Department[] ds = new Department[numOfDepartments*2];
for(int i=0; i<numOfDepartments;i++){
ds[i]=departments[i];
}
}
departments[numOfDepartments++] = new Department(name);
}
public Iterator<?> createIterator() {
// TODO Auto-generated method stub
return new ComputerCollegeIterator<>(departments,numOfDepartments);
}
}
public class InfoCollegeIterator implements Iterator<Object>{
List<Department> departments;
int index = -1;
public InfoCollegeIterator(List<Department> departments) {
super();
this.departments = departments;
}
@Override
public boolean hasNext() {
// TODO Auto-generated method stub
if (index >= departments.size()-1) {
return false;
}
index++;
return true;
}
@Override
public Object next() {
// TODO Auto-generated method stub
return departments.get(index);
}
}
public class InfoCollege extends College{
List<Department> departments;
public InfoCollege(String name) {
super(name);
// TODO Auto-generated constructor stub
departments = new ArrayList<Department>();
addDepartment("網(wǎng)安");
}
public void addDepartment(String name) {
departments.add(new Department(name));
}
public Iterator<?> createIterator() {
// TODO Auto-generated method stub
return new InfoCollegeIterator(departments);
}
}
public class OutputImpl {
private List<College> colleges;
public OutputImpl(List<College> colleges) {
super();
this.colleges = colleges;
}
public void printCollege(){
for(College college :colleges){
System.out.println("======="+college.getName()+"======");
printDepartment(college.createIterator());
}
}
public void printDepartment(Iterator<?> iterator){
while (iterator.hasNext()) {
Department next = (Department) iterator.next();
System.out.println(next.getName());
}
}
}
public class Client {
public static void main(String[] args) {
College compute = new ComputerCollege("計(jì)算機(jī)學(xué)院");
College info = new InfoCollege("信息學(xué)院");
@SuppressWarnings("serial")
OutputImpl output = new OutputImpl(new ArrayList<College>(){
{add(compute);add(info);}});
output.printCollege();
}
}
迭代器模式主要作用是提供統(tǒng)一的遍歷接口，此例中computeCollege持有Department的形式是數(shù)組，而InfoCollege持有的形式是List，為這兩種形式提供統(tǒng)一的遍歷方式。
迭代器模式的注意事項(xiàng)和細(xì)節(jié)
?優(yōu)點(diǎn)
- 提供一個(gè)統(tǒng)一的方法遍歷對(duì)象，客戶不用再考慮聚合的類型，使用-種方法就可以遍歷對(duì)象了。
- 隱藏了聚合的內(nèi)部結(jié)構(gòu)，客戶端要遍歷聚合的時(shí)候只能取到迭代器，而不會(huì)知道聚合的具體組成。
- 提供了一種設(shè)計(jì)思想，就是一個(gè)類應(yīng)該只有一一個(gè)引起變化的原因(叫做單一責(zé)任原則)。在聚合類中，我們把迭代器分開，就是要把管理對(duì)象集合和遍歷對(duì)象集合的責(zé)任分開，這樣- -來集合改變的話，只影響到聚合對(duì)象。而如果遍歷方式改變的話，只影響到了迭代器。
- 當(dāng)要展示一組相似對(duì)象，或者遍歷一組相同對(duì)象時(shí)使用，適合使用迭代器模式
?缺點(diǎn)
每個(gè)聚合對(duì)象都要-一個(gè)迭代器， 會(huì)生成多個(gè)迭代器不好管理類
