我需要能够在一段时间内将对象存储在数据结构中，然后能够获取特定时间数据的快照。我有一些解决方案，但我想知道是否有更优化的方法来做到这一点。我想可能有一种方法可以优化存储机制，以便不需要迭代每个元素来获取当时的快照。
在实际情况中，可以随时更新数据。数据结构中的对象由Integers索引。我用Map快速模拟了一些东西来说明我想要做的事情（arraylist是对象更新的历史）。
public class TimeStoreTest {
static SimpleDateFormat df = new SimpleDateFormat("yy-MM-dd HH:mm:ss");
public static void main(String[] args) throws Exception {
Date[] dates = new Date[] { df.parse("2014-01-23 10:30:30"),
df.parse("14-01-23 10:33:30"), df.parse("14-01-23 10:36:30"),
df.parse("14-01-23 10:39:30"), df.parse("14-01-23 10:42:30") };
Map<Integer, List<SimpleObject>> timeStore = new HashMap<Integer, List<SimpleObject>>();
for (int i = 0; i < 10; i += 2) {
List<SimpleObject> objs = new ArrayList<SimpleObject>();
objs.add(new SimpleObject(i, dates[0]));
timeStore.put(i, objs);
}
for (int i = 1; i < 10; i += 2) {
List<SimpleObject> objs = new ArrayList<SimpleObject>();
objs.add(new SimpleObject(i, dates[1]));
timeStore.put(i, objs);
}
for (int i = 0; i < 10; i += 2) {
List<SimpleObject> objs = timeStore.get(i);
objs.add(new SimpleObject(i, dates[2]));
}
for (int i = 1; i < 10; i += 2) {
List<SimpleObject> objs = timeStore.get(i);
objs.add(new SimpleObject(i, dates[3]));
}
for (int i = 5; i < 10; i++) {
List<SimpleObject> objs = timeStore.get(i);
objs.add(new SimpleObject(i, dates[4]));
}
System.out.println("Current Data Store: ");
for (Entry<Integer, List<SimpleObject>> entry : timeStore.entrySet()) {
System.out.println(entry);
}
Date queryDate = df.parse("14-01-23 10:31:00");
System.out.println("Query for object snapshot @ " + queryDate);
for (SimpleObject obj : getObjectsAtTime(queryDate, timeStore)) {
System.out.println(obj);
}
queryDate = df.parse("14-01-23 10:34:30");
System.out.println("Query for object snapshot @ " + queryDate);
for (SimpleObject obj : getObjectsAtTime(queryDate, timeStore)) {
System.out.println(obj);
}
queryDate = df.parse("14-01-23 10:39:30");
System.out.println("Query for object snapshot @ " + queryDate);
for (SimpleObject obj : getObjectsAtTime(queryDate, timeStore)) {
System.out.println(obj);
}
queryDate = df.parse("14-01-23 10:42:00");
System.out.println("Query for object snapshot @ " + queryDate);
for (SimpleObject obj : getObjectsAtTime(queryDate, timeStore)) {
System.out.println(obj);
}
queryDate = df.parse("14-01-23 10:45:00");
System.out.println("Query for object snapshot @ " + queryDate);
for (SimpleObject obj : getObjectsAtTime(queryDate, timeStore)) {
System.out.println(obj);
}
}
public static List<SimpleObject> getObjectsAtTime(Date date,
Map<Integer, List<SimpleObject>> store) {
List<SimpleObject> objectsAtTime = new ArrayList<SimpleObject>();
for (Entry<Integer, List<SimpleObject>> entry : store.entrySet()) {
SimpleObject previousObject = null;
List<SimpleObject> list = entry.getValue();
for (int i = 0; i < list.size(); i++) {
SimpleObject currentObject = list.get(i);
if (previousObject != null) {
if (date.after(previousObject.updatedTime)
&& date.before(currentObject.updatedTime)) {
objectsAtTime.add(previousObject);
break;
} else if (currentObject.updatedTime.equals(date)) {
objectsAtTime.add(currentObject);
break;
} else if (i == list.size() - 1
&& date.after(previousObject.updatedTime)) {
objectsAtTime.add(currentObject);
}
}
previousObject = currentObject;
}
}
return objectsAtTime;
}
}
class SimpleObject implements Comparable {
int id;
Date updatedTime;
public SimpleObject(int id, Date updatedTime) {
this.id = id;
this.updatedTime = new Date(updatedTime.getTime());
}
@Override
public String toString() {
return "(object id=" + id + "; t="
+ TimeStoreTest.df.format(updatedTime) + ")";
}
@Override
public int compareTo(Object object) {
// TODO Auto-generated method stub
return updatedTime.compareTo(((SimpleObject) object).updatedTime);
}
}
结果：
当前数据存储：
0=[(object id=0; t=14-01-23 10:30:30), (object id=0; t=14-01-23 10:36:30)]
1=[(object id=1; t=14-01-23 10:33:30), (object id=1; t=14-01-23 10:39:30)]
2=[(object id=2; t=14-01-23 10:30:30), (object id=2; t=14-01-23 10:36:30)]
3=[(object id=3; t=14-01-23 10:33:30), (object id=3; t=14-01-23 10:39:30)]
4=[(object id=4; t=14-01-23 10:30:30), (object id=4; t=14-01-23 10:36:30)]
5=[(object id=5; t=14-01-23 10:33:30), (object id=5; t=14-01-23 10:39:30), (object id=5; t=14-01-23 10:42:30)]
6=[(object id=6; t=14-01-23 10:30:30), (object id=6; t=14-01-23 10:36:30), (object id=6; t=14-01-23 10:42:30)]
7=[(object id=7; t=14-01-23 10:33:30), (object id=7; t=14-01-23 10:39:30), (object id=7; t=14-01-23 10:42:30)]
8=[(object id=8; t=14-01-23 10:30:30), (object id=8; t=14-01-23 10:36:30), (object id=8; t=14-01-23 10:42:30)]
9=[(object id=9; t=14-01-23 10:33:30), (object id=9; t=14-01-23 10:39:30), (object id=9; t=14-01-23 10:42:30)]
Query for object snapshot @ Thu Jan 23 10:31:00 PST 2014
(object id=0; t=14-01-23 10:30:30)
(object id=2; t=14-01-23 10:30:30)
(object id=4; t=14-01-23 10:30:30)
(object id=6; t=14-01-23 10:30:30)
(object id=8; t=14-01-23 10:30:30)
Query for object snapshot @ Thu Jan 23 10:34:30 PST 2014
(object id=0; t=14-01-23 10:30:30)
(object id=1; t=14-01-23 10:33:30)
(object id=2; t=14-01-23 10:30:30)
(object id=3; t=14-01-23 10:33:30)
(object id=4; t=14-01-23 10:30:30)
(object id=5; t=14-01-23 10:33:30)
(object id=6; t=14-01-23 10:30:30)
(object id=7; t=14-01-23 10:33:30)
(object id=8; t=14-01-23 10:30:30)
(object id=9; t=14-01-23 10:33:30)
Query for object snapshot @ Thu Jan 23 10:39:30 PST 2014
(object id=0; t=14-01-23 10:36:30)
(object id=1; t=14-01-23 10:39:30)
(object id=2; t=14-01-23 10:36:30)
(object id=3; t=14-01-23 10:39:30)
(object id=4; t=14-01-23 10:36:30)
(object id=5; t=14-01-23 10:39:30)
(object id=6; t=14-01-23 10:36:30)
(object id=7; t=14-01-23 10:39:30)
(object id=8; t=14-01-23 10:36:30)
(object id=9; t=14-01-23 10:39:30)
Query for object snapshot @ Thu Jan 23 10:42:00 PST 2014
(object id=0; t=14-01-23 10:36:30)
(object id=1; t=14-01-23 10:39:30)
(object id=2; t=14-01-23 10:36:30)
(object id=3; t=14-01-23 10:39:30)
(object id=4; t=14-01-23 10:36:30)
(object id=5; t=14-01-23 10:39:30)
(object id=6; t=14-01-23 10:36:30)
(object id=7; t=14-01-23 10:39:30)
(object id=8; t=14-01-23 10:36:30)
(object id=9; t=14-01-23 10:39:30)
Query for object snapshot @ Thu Jan 23 10:45:00 PST 2014
(object id=0; t=14-01-23 10:36:30)
(object id=1; t=14-01-23 10:39:30)
(object id=2; t=14-01-23 10:36:30)
(object id=3; t=14-01-23 10:39:30)
(object id=4; t=14-01-23 10:36:30)
(object id=5; t=14-01-23 10:42:30)
(object id=6; t=14-01-23 10:42:30)
(object id=7; t=14-01-23 10:42:30)
(object id=8; t=14-01-23 10:42:30)
(object id=9; t=14-01-23 10:42:30)
所以，你可以看到它有效，但我正试图找到提高效率的方法。
答案 0 :(得分：1)
使用NavigableMap可能会让您的生活更轻松。请查看TreeMap#floorEntry()处的示例，该示例与您的getObjectsAtTime方法类似。
它还有其他方法，例如subMap也可能很方便。
答案 1 :(得分：1)
好像你可以使用TreeMap然后调用subMap来查询你想要的日期范围
import java.text.SimpleDateFormat;
import java.util.Date;
import java.util.NavigableMap;
import java.util.TreeMap;
public class KeyRange {
static SimpleDateFormat df = new SimpleDateFormat("yy-MM-dd HH:mm:ss");
public static void main(String[] args) throws Exception {
Date[] dates = new Date[]{df.parse("2014-01-23 10:30:30"),
df.parse("14-01-23 10:33:30"), df.parse("14-01-23 10:36:30"),
df.parse("14-01-23 10:39:30"), df.parse("14-01-23 10:42:30")};
/**
* use a treeMap, because it has sorted keys and it can be queried with from -> to values
*/
TreeMap<Date, Object> data = new TreeMap<>();
for(Date d : dates){
data.put(d, null);
}
/**
* use the subMap function to "query" the map by a range of sorted keys...this would be your query time span
*/
NavigableMap<Date, Object> queryResult = data.subMap(df.parse("14-01-23 10:33:30"), true, df.parse("14-01-23 10:39:30"), false);
System.out.println(queryResult);
}
}
上面的结果就是这个
{Thu 1月23日10:33:30 EST 2014 = null，Thu 1月23日10:36:30 EST 2014 = null}
如果您希望它按降序排序，则将.descendingMap（）添加到子图方法的末尾
答案 2 :(得分：0)
我最终选择了一个区间树（http://en.wikipedia.org/wiki/Interval_tree），每个段都被映射到当时活动的对象的HashMap。
我为我的用法修改了这段代码：