在Spring Data MongoDB中，可以使用如下方法更新文档：

1. save – 如果_id存在则更新，否则插入。更新整个文档。
2. updateFirst – 更新查询出的第一个文档。
3. updateMulti – 更新查询出的所有文档。
4. upsert – 如果没有查询出文档，则会创建一个新文档。
5. findAndModify – 和updateMulti相同，但是它有一个额外的选项可以返回更新前或更新后的文档。

1. saveOrUpdate – 例1

假设下面的json数据插入到MongoDB。

{
    "_id" : ObjectId("id"),
    "ic" : "1001",
    "name" : "appleA",
    "age" : 20,
    "createdDate" : ISODate("2013-04-06T23:17:35.530Z")
}

查询文档，并使用save()方法修改和更新。

Query query = new Query();
query.addCriteria(Criteria.where("name").is("appleA"));

User userTest1 = mongoOperation.findOne(query, User.class);

System.out.println("userTest1 - " + userTest1);

//modify and update with save()
userTest1.setAge(99);
mongoOperation.save(userTest1);

//get the updated object again
User userTest1_1 = mongoOperation.findOne(query, User.class);

System.out.println("userTest1_1 - " + userTest1_1);

输出：

userTest1 - User [id=id, ic=1001, name=appleA, age=20, createdDate=Sat Apr 06 23:17:35 MYT 2013]
userTest1_1 - User [id=id, ic=1001, name=appleA, age=99, createdDate=Sat Apr 06 23:17:35 MYT 2013]

1. saveOrUpdate – 例2

这是一个失败的例子，请仔细阅读，这是一个相当普遍的错误。

假设下面的json数据插入到MongoDB。

{
    "_id" : ObjectId("id"),
    "ic" : "1002",
    "name" : "appleB",
    "age" : 20,
    "createdDate" : ISODate("2013-04-06T15:22:34.530Z")
}

使用Query获取只包含name字段的文档，返回的User对象的age、ic和createDate字段都为null，如果要修改age字段并更新，它会覆盖所有值而不是更新age字段。

Query query = new Query();
query.addCriteria(Criteria.where("name").is("appleB"));
query.fields().include("name");

User userTest2 = mongoOperation.findOne(query, User.class);
System.out.println("userTest2 - " + userTest2);

userTest2.setAge(99);

mongoOperation.save(userTest2);

// ooppss, you just override everything, it caused ic=null and
// createdDate=null

Query query1 = new Query();
query1.addCriteria(Criteria.where("name").is("appleB"));

User userTest2_1 = mongoOperation.findOne(query1, User.class);
System.out.println("userTest2_1 - " + userTest2_1);

输出：

userTest2 - User [id=51603dba3004d7fffc202391, ic=null, name=appleB, age=0, createdDate=null]
userTest2_1 - User [id=51603dba3004d7fffc202391, ic=null, name=appleB, age=99, createdDate=null]

save()方法之后，age字段被正确更新了，但是ic和createdDate字段都是null，整个user对象都被更新了。为了只更新某个字段而不是整个对象，可以使用updateFirst()或者updateMulti()，而不是save()。

3. updateFirst

更新符合查询条件的第一个文档。本例中，只有age字段被更新。

{
    "_id" : ObjectId("id"),
    "ic" : "1003",
    "name" : "appleC",
    "age" : 20,
    "createdDate" : ISODate("2013-04-06T23:22:34.530Z")
}

//returns only 'name' field
Query query = new Query();
query.addCriteria(Criteria.where("name").is("appleC"));
query.fields().include("name");

User userTest3 = mongoOperation.findOne(query, User.class);
System.out.println("userTest3 - " + userTest3);

Update update = new Update();
update.set("age", 100);

mongoOperation.updateFirst(query, update, User.class);

//returns everything
Query query1 = new Query();
query1.addCriteria(Criteria.where("name").is("appleC"));

User userTest3_1 = mongoOperation.findOne(query1, User.class);
System.out.println("userTest3_1 - " + userTest3_1);

输出：

userTest3 - User [id=id, ic=null, name=appleC, age=0, createdDate=null]
userTest3_1 - User [id=id, ic=1003, name=appleC, age=100, createdDate=Sat Apr 06 23:22:34 MYT 2013]

4. updateMulti

更新符合查询条件的所有文档。

{
    "_id" : ObjectId("id"),
    "ic" : "1004",
    "name" : "appleD",
    "age" : 20,
    "createdDate" : ISODate("2013-04-06T15:22:34.530Z")
}
{
    "_id" : ObjectId("id"),
    "ic" : "1005",
    "name" : "appleE",
    "age" : 20,
    "createdDate" : ISODate("2013-04-06T15:22:34.530Z")
}

//show the use of $or operator
Query query = new Query();
query.addCriteria(Criteria
        .where("name").exists(true)
        .orOperator(Criteria.where("name").is("appleD"),
                Criteria.where("name").is("appleE")));
Update update = new Update();

//update age to 11
update.set("age", 11);

//remove the createdDate field
update.unset("createdDate");

// if use updateFirst, it will update 1004 only.
// mongoOperation.updateFirst(query4, update4, User.class);

// update all matched, both 1004 and 1005
mongoOperation.updateMulti(query, update, User.class);

System.out.println(query.toString());

List<User> usersTest4 = mongoOperation.find(query4, User.class);

for (User userTest4 : usersTest4) {
    System.out.println("userTest4 - " + userTest4);
}

输出：

Query: { "name" : { "$exists" : true} ,
    "$or" : [ { "name" : "appleD"} , { "name" : "appleE"}]}, Fields: null, Sort: null

userTest4 - User [id=id, ic=1004, name=appleD, age=11, createdDate=null]
userTest4 - User [id=id, ic=1005, name=appleE, age=11, createdDate=null]

5. upsert

如果查询到符合的文档就更新，如果没有就根据查询条件创建一个新的文档，就像findAndModifyElseCreate()。

{
    //no data
}

//search a document that doesn't exist
Query query = new Query();
query.addCriteria(Criteria.where("name").is("appleZ"));

Update update = new Update();
update.set("age", 21);

mongoOperation.upsert(query, update, User.class);

User userTest5 = mongoOperation.findOne(query, User.class);
System.out.println("userTest5 - " + userTest5);

输出：

userTest5 - User [id=id, ic=null, name=appleZ, age=21, createdDate=null]

6. findAndModify

使用一个操作查询、修改并获得更新后的对象。

{
    "_id" : ObjectId("id"),
    "ic" : "1006",
    "name" : "appleF",
    "age" : 20,
    "createdDate" : ISODate("2013-04-07T13:11:34.530Z")
}

Query query6 = new Query();
query6.addCriteria(Criteria.where("name").is("appleF"));

Update update6 = new Update();
update6.set("age", 101);
update6.set("ic", 1111);

//FindAndModifyOptions().returnNew(true) = newly updated document
//FindAndModifyOptions().returnNew(false) = old document (not update yet)
User userTest6 = mongoOperation.findAndModify(
        query6, update6,
        new FindAndModifyOptions().returnNew(true), User.class);
System.out.println("userTest6 - " + userTest6);

输出：

userTest6 - User [id=id, ic=1111, name=appleF, age=101, createdDate=Sun Apr 07 13:11:34 MYT 2013]

7. 完整实例

import java.util.ArrayList;
import java.util.Date;
import java.util.List;

import org.springframework.context.ApplicationContext;
import org.springframework.context.annotation.AnnotationConfigApplicationContext;
import org.springframework.data.mongodb.core.FindAndModifyOptions;
import org.springframework.data.mongodb.core.MongoOperations;
import org.springframework.data.mongodb.core.query.Criteria;
import org.springframework.data.mongodb.core.query.Query;
import org.springframework.data.mongodb.core.query.Update;

import com.mkyong.config.SpringMongoConfig;
import com.mkyong.model.User;

public class UpdateApp {

    public static void main(String[] args) {
        // For Annotation
        ApplicationContext ctx =
            new AnnotationConfigApplicationContext(SpringMongoConfig.class);
        MongoOperations mongoOperation =
            (MongoOperations) ctx.getBean("mongoTemplate");

        // insert 6 users for testing
        List<User> users = new ArrayList<User>();

        User user1 = new User("1001", "appleA", 20, new Date());
        User user2 = new User("1002", "appleB", 20, new Date());
        User user3 = new User("1003", "appleC", 20, new Date());
        User user4 = new User("1004", "appleD", 20, new Date());
        User user5 = new User("1005", "appleE", 20, new Date());
        User user6 = new User("1006", "appleF", 20, new Date());
        users.add(user1);
        users.add(user2);
        users.add(user3);
        users.add(user4);
        users.add(user5);
        users.add(user6);
        mongoOperation.insert(users, User.class);

        // Case 1 ... find and update
        System.out.println("Case 1");

        Query query1 = new Query();
        query1.addCriteria(Criteria.where("name").is("appleA"));

        User userTest1 = mongoOperation.findOne(query1, User.class);

        System.out.println("userTest1 - " + userTest1);

        userTest1.setAge(99);
        mongoOperation.save(userTest1);

        User userTest1_1 = mongoOperation.findOne(query1, User.class);

        System.out.println("userTest1_1 - " + userTest1_1);

        // Case 2 ... select single field only
        System.out.println("\nCase 2");

        Query query2 = new Query();
        query2.addCriteria(Criteria.where("name").is("appleB"));
        query2.fields().include("name");

        User userTest2 = mongoOperation.findOne(query2, User.class);
        System.out.println("userTest2 - " + userTest2);

        userTest2.setAge(99);

        mongoOperation.save(userTest2);

        // ooppss, you just override everything, it caused ic=null and
        // createdDate=null

        Query query2_1 = new Query();
        query2_1.addCriteria(Criteria.where("name").is("appleB"));

        User userTest2_1 = mongoOperation.findOne(query2_1, User.class);
        System.out.println("userTest2_1 - " + userTest2_1);

        System.out.println("\nCase 3");
        Query query3 = new Query();
        query3.addCriteria(Criteria.where("name").is("appleC"));
        query3.fields().include("name");

        User userTest3 = mongoOperation.findOne(query3, User.class);
        System.out.println("userTest3 - " + userTest3);

        Update update3 = new Update();
        update3.set("age", 100);

        mongoOperation.updateFirst(query3, update3, User.class);

        Query query3_1 = new Query();
        query3_1.addCriteria(Criteria.where("name").is("appleC"));

        User userTest3_1 = mongoOperation.findOne(query3_1, User.class);
        System.out.println("userTest3_1 - " + userTest3_1);

        System.out.println("\nCase 4");
        Query query4 = new Query();
        query4.addCriteria(Criteria
                .where("name")
                .exists(true)
                .orOperator(Criteria.where("name").is("appleD"),
                        Criteria.where("name").is("appleE")));
        Update update4 = new Update();
        update4.set("age", 11);
        update4.unset("createdDate");

        // update 1004 only.
        // mongoOperation.updateFirst(query4, update4, User.class);

        // update all matched
        mongoOperation.updateMulti(query4, update4, User.class);

        System.out.println(query4.toString());

        List<User> usersTest4 = mongoOperation.find(query4, User.class);

        for (User userTest4 : usersTest4) {
            System.out.println("userTest4 - " + userTest4);
        }

        System.out.println("\nCase 5");
        Query query5 = new Query();
        query5.addCriteria(Criteria.where("name").is("appleZ"));

        Update update5 = new Update();
        update5.set("age", 21);

        mongoOperation.upsert(query5, update5, User.class);

        User userTest5 = mongoOperation.findOne(query5, User.class);
        System.out.println("userTest5 - " + userTest5);

        System.out.println("\nCase 6");
        Query query6 = new Query();
        query6.addCriteria(Criteria.where("name").is("appleF"));

        Update update6 = new Update();
        update6.set("age", 101);
        update6.set("ic", 1111);

        User userTest6 = mongoOperation.findAndModify(query6, update6,
                new FindAndModifyOptions().returnNew(true), User.class);
        System.out.println("userTest6 - " + userTest6);

        mongoOperation.dropCollection(User.class);

    }

}

输出：

Case 1
userTest1 - User [id=id, ic=1001, name=appleA, age=20, createdDate=Sun Apr 07 13:22:48 MYT 2013]
userTest1_1 - User [id=id, ic=1001, name=appleA, age=99, createdDate=Sun Apr 07 13:22:48 MYT 2013]

Case 2
userTest2 - User [id=id, ic=null, name=appleB, age=0, createdDate=null]
userTest2_1 - User [id=id, ic=null, name=appleB, age=99, createdDate=null]

Case 3
userTest3 - User [id=id, ic=null, name=appleC, age=0, createdDate=null]
userTest3_1 - User [id=id, ic=1003, name=appleC, age=100, createdDate=Sun Apr 07 13:22:48 MYT 2013]

Case 4
Query: { "name" : { "$exists" : true} , "$or" : [ { "name" : "appleD"} , { "name" : "appleE"}]}, Fields: null, Sort: null
userTest4 - User [id=id, ic=1004, name=appleD, age=11, createdDate=null]
userTest4 - User [id=id, ic=1005, name=appleE, age=11, createdDate=null]

Case 5
userTest5 - User [id=id, ic=null, name=appleZ, age=21, createdDate=null]

Case 6
userTest6 - User [id=id, ic=1006, name=appleF, age=20, createdDate=Sun Apr 07 13:22:48 MYT 2013]