Stream闪亮登场
一. Stream(流)是什么,干什么
Stream是一类用于替代对集合操作的工具类+Lambda式编程,他可以替代现有的遍历、过滤、求和、求最值、排序、转换等
二. Stream操作方式
- 并行方式parallelStream
- 顺序方式Stream
三. Stream优势
- Lambda 可有效减少冗余代码,减少开发工作量
- 内置对集合List、Map的多种操作方式,含基本数据类型处理
- 并行Stream有效率优势(内置多线程)
四. Stream(流)的基本使用
- 遍历forEach
@Test
public void stream() {
//操作List
List<Map<String, String>> mapList = new ArrayList() {
{
Map<String, String> m = new HashMap();
m.put("a", "1");
Map<String, String> m2 = new HashMap();
m2.put("b", "2");
add(m);
add(m2);
}
};
mapList.stream().forEach(item-> System.out.println(item));
//操作Map
Map<String,Object> mp = new HashMap(){
{
put("a","1");
put("b","2");
put("c","3");
put("d","4");
}
};
mp.keySet().stream().forEachOrdered(item-> System.out.println(mp.get(item)));
}
- 过滤filter
List<Integer> mapList = new ArrayList() {
{
add(1);
add(10);
add(12);
add(33);
add(99);
}
};
//mapList.stream().forEach(item-> System.out.println(item));
mapList = mapList.stream().filter(item->{
return item>30;
}).collect(Collectors.toList());
System.out.println(mapList);
- 转换map和极值
@Test public void trans(){ List<Person> ps = new ArrayList<Person>(){ { Person p1 = new Person(); p1.setAge(11); p1.setName("张强"); Person p2 = new Person(); p2.setAge(17); p2.setName("李思"); Person p3 = new Person(); p3.setAge(20); p3.setName("John"); add(p1); add(p2); add(p3); } }; //取出所有age字段为一个List List<Integer> sumAge = ps.stream().map(Person::getAge).collect(Collectors.toList()); System.out.println(sumAge); //取出age最大的那 Integer maxAge =sumAge.stream().max(Integer::compare).get(); System.out.println(maxAge); } class Person{ private String name; private Integer age; public String getName() { return name; } public void setName(String name) { this.name = name; } public Integer getAge() { return age; } public void setAge(Integer age) { this.age = age; }
}
#### 五. Stream(流)的效率
+ 模拟非耗时简单业务逻辑
class Person{
private String name;
private int age;
private Date joinDate;
private String label;
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
public int getAge() {
return age;
}
public void setAge(int age) {
this.age = age;
}
public Date getJoinDate() {
return joinDate;
}
public void setJoinDate(Date joinDate) {
this.joinDate = joinDate;
}
public String getLabel() {
return label;
}
public void setLabel(String label) {
this.label = label;
}
public class DataLoopTest {
private static final Logger LOG= LoggerFactory.getLogger(DataLoopTest.class);
private static final List<Person> persons = new ArrayList<>();
static {
for(int i=0;i<=1000000;i++){
Person p = new Person();
p.setAge(i);
p.setName("zhangSan");
p.setJoinDate(new Date());
persons.add(p);
}
}
/**
* for 循环耗时 ===> 1.988
* for 循环耗时 ===> 2.198
* for 循环耗时 ===> 1.978
*
*/
@Test
public void forTest(){
Instant date_start = Instant.now();
int personSize = persons.size();
for(int i=0;i<personSize;i++){
persons.get(i).setLabel(persons.get(i).getName().concat("-"+persons.get(i).getAge()).concat("-"+persons.get(i).getJoinDate().getTime()));
}
Instant date_end = Instant.now();
LOG.info("for 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
/**
* forEach 循环耗时 ===> 1.607
* forEach 循环耗时 ===> 2.242
* forEach 循环耗时 ===> 1.875
*/
@Test
public void forEach(){
Instant date_start = Instant.now();
for(Person p:persons){
p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime()));
}
Instant date_end = Instant.now();
LOG.info("forEach 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
/**
* streamForeach 循环耗时 ===> 1.972
* streamForeach 循环耗时 ===> 1.969
* streamForeach 循环耗时 ===> 2.125
*/
@Test
public void streamForeach(){
Instant date_start = Instant.now();
persons.stream().forEach(p->p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime())));
Instant date_end = Instant.now();
LOG.info("streamForeach 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
/**
* parallelStreamForeach 循环耗时 ===> 1.897
* parallelStreamForeach 循环耗时 ===> 1.942
* parallelStreamForeach 循环耗时 ===> 1.642
*/
@Test
public void parallelStreamForeach(){
Instant date_start = Instant.now();
persons.parallelStream().forEach(p->p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime())));
Instant date_end = Instant.now();
LOG.info("parallelStreamForeach 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
}
+ 模拟耗时简单业务逻辑
public class DataLoopBlockTest {
private static final Logger LOG= LoggerFactory.getLogger(DataLoopTest.class);
private static final List<Person> persons = new ArrayList<>();
static {
for(int i=0;i<=100000;i++){
Person p = new Person();
p.setAge(i);
p.setName("zhangSan");
p.setJoinDate(new Date());
persons.add(p);
}
}
/**
* for 循环耗时 ===> 101.385
* for 循环耗时 ===> 102.161
* for 循环耗时 ===> 101.472
*
*/
@Test
public void forTest(){
Instant date_start = Instant.now();
int personSize = persons.size();
for(int i=0;i<personSize;i++){
try {
Thread.sleep(1);
persons.get(i).setLabel(persons.get(i).getName().concat("-"+persons.get(i).getAge()).concat("-"+persons.get(i).getJoinDate().getTime()));
}catch (Exception e){
e.printStackTrace();
}
}
Instant date_end = Instant.now();
LOG.info("for 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
/**
* forEach 循环耗时 ===> 101.027
* forEach 循环耗时 ===> 102.488
* forEach 循环耗时 ===> 101.608
*/
@Test
public void forEach(){
Instant date_start = Instant.now();
for(Person p:persons){
try {
Thread.sleep(1);
p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime()));
}catch (Exception e){
e.printStackTrace();
}
}
Instant date_end = Instant.now();
LOG.info("forEach 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
/**
* streamForeach 循环耗时 ===> 103.246
* streamForeach 循环耗时 ===> 101.128
* streamForeach 循环耗时 ===> 102.615
*/
@Test
public void streamForeach(){
Instant date_start = Instant.now();
//persons.stream().forEach(p->p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime())));
persons.stream().forEach(p->{
try {
Thread.sleep(1);
p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime()));
}catch (Exception e){
e.printStackTrace();
}
});
Instant date_end = Instant.now();
LOG.info("streamForeach 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
}
/**
* parallelStreamForeach 循环耗时 ===> 51.391
* parallelStreamForeach 循环耗时 ===> 53.509
* parallelStreamForeach 循环耗时 ===> 50.831
*/
@Test
public void parallelStreamForeach(){
Instant date_start = Instant.now();
//persons.parallelStream().forEach(p->p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime())));
persons.parallelStream().forEach(p->{
try {
Thread.sleep(1);
p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime()));
}catch (Exception e){
e.printStackTrace();
}
});
Instant date_end = Instant.now();
LOG.info("parallelStreamForeach 循环耗时 ===> {}", Duration.between(date_start,date_end).toMillis()/1000.0);
//LOG.info("
===> {}",JSON.toJSONString(persons.get(10000)));
}
}
可以看到在百万数据下做简单数据循环处理,对于普通for(forforeach)循环或stream(并行、非并行)下,几者的效率差异并不明显,
注意: 在百万数据下,普通for、foreach循环处理可能比stream的方式快许多,对于这点效率的损耗,其实lambda表达式对代码的简化更大!
另外,在并行流的循环下速度提升了一倍之多,当单个循环耗时较多时,会拉大与前几者的循环效率
(以上测试仅对于循环而言,其他类型业务处理,比如排序、求和、最大值等未做测试,个人猜测与以上测试结果相似)
#### 六. Stream(流)注意项
+ 并行stream不是线程安全的,当对循坏外部统一对象进行读写时候会造成意想不到的错误,这需要留意
+ 因stream总是惰性的,原对象是不可以被修改的,在集合处理完成后需要将处理结果放入一个新的集合容器内
+ 普通循环与stream(非并行)循环,在处理处理数据量比较大的时候效率是一致的,推荐使用stream的形式
+ 对于List删除操作,目前只提供了removeIf方法来实现,并不能使用并行方式
+ 对于lambda表达式的写法
- 当表达式内只有一个返回boolean类型的语句时候语句是可以简写的,例如:
persons.parallelStream().forEach(p->p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime())));
- 当表达式内会有一些复杂处理逻辑时需要加上大括号,这与初始化List参数方式大致一致
persons.parallelStream().forEach(p->{
try {
Thread.sleep(1);
p.setLabel(p.getName().concat("-"+p.getAge()).concat("-"+p.getJoinDate().getTime()));
}catch (Exception e){
e.printStackTrace();
}
});
#### 七. stream&Lambda表达式常用api方法
+ 流到流之间的转换类
- filter(过滤)
- map(映射转换)
- mapTo[Int|Long|Double] (到基本类型流的转换)
- flatMap(流展开合并)
- flatMapTo[Int|Long|Double]
- sorted(排序)
- distinct(不重复值)
- peek(执行某种操作,流不变,可用于调试)
- limit(限制到指定元素数量)
- skip(跳过若干元素)
+ 流到终值的转换类
- toArray(转为数组)
- reduce(推导结果)
- collect(聚合结果)
- min(最小值)
- max(最大值)
- count (元素个数)
- anyMatch (任一匹配)
- allMatch(所有都匹配)
- noneMatch(一个都不匹配)
- findFirst(选择首元素)
- findAny(任选一元素)
+ 直接遍历类
- forEach (不保证顺序遍历,比如并行流)
- forEachOrdered(顺序遍历)
+ 构造流类
- empty (构造空流)
- of (单个元素的流及多元素顺序流)
- iterate (无限长度的有序顺序流)
- generate (将数据提供器转换成无限非有序的顺序流)
- concat (流的连接)
- Builder (用于构造流的Builder对象)