MongoDB - 聚合阶段 $match、$sort、$limit

1. $match 聚合阶段

$match 接受一个指定查询条件的文档。

$match 阶段语法：

{ $match: { <query> } }

$match 查询语法与读取操作查询语法相同，即 $match 不接受原始聚合表达式。要在 $match 中包含聚合表达式，请使用 $expr 查询表达式：

{ $match: { $expr: { <aggregation expression> } } }

尽可能早地将 $match 放在聚合管道中，由于 $match 限制了聚合管道中的文档总数，因此早期的 $match 操作会最大限度地减少管道中的处理量。

1. 构造测试数据

db.articles.drop()db.articles.insertMany([{"_id": ObjectId("512bc95fe835e68f199c8686"),"author": "dave","score": 80,"views": 100},{"_id": ObjectId("512bc962e835e68f199c8687"),"author": "dave","score": 85,"views": 521},{"_id": ObjectId("55f5a192d4bede9ac365b257"),"author": "ahn","score": 60,"views": 1000},{"_id": ObjectId("55f5a192d4bede9ac365b258"),"author": "li","score": 55,"views": 5000},{"_id": ObjectId("55f5a1d3d4bede9ac365b259"),"author": "annT","score": 60,"views": 50},{"_id": ObjectId("55f5a1d3d4bede9ac365b25a"),"author": "li","score": 94,"views": 999},{"_id": ObjectId("55f5a1d3d4bede9ac365b25b"),"author": "ty","score": 95,"views": 1000}
])

2. $match 示例

使用 $match 来执行简易等值匹配，$match 会选择 author 字段等于 dave 的文档，而聚合返回以下内容：

db.articles.aggregate([ { $match : { author : "dave" } } ]
);

// 1
{"_id": ObjectId("512bc95fe835e68f199c8686"),"author": "dave","score": 80,"views": 100
}// 2
{"_id": ObjectId("512bc962e835e68f199c8687"),"author": "dave","score": 85,"views": 521
}

SpringBoot 整合 MongoDB 实现：

@SpringBootTest
@RunWith(SpringRunner.class)
public class BeanLoadServiceTest {@Autowiredprivate MongoTemplate mongoTemplate;@Testpublic void aggregateTest() {// $match阶段Criteria criteria = Criteria.where("author").is("dave");MatchOperation match = Aggregation.match(criteria);Aggregation aggregation = Aggregation.newAggregation(match);// 执行聚合管道操作AggregationResults<Article> results= mongoTemplate.aggregate(aggregation, Article.class, Article.class);List<Article> mappedResults = results.getMappedResults();// 打印结果mappedResults.forEach(System.out::println);//Article(id=512bc95fe835e68f199c8686, author=dave, score=80, views=100)//Article(id=512bc962e835e68f199c8687, author=dave, score=85, views=521)}
}

这里输出文档直接使用了Article.class，可以重新定义实体类接收输出文档的字段：

@Data
public class AggregationResult {@Idprivate String id;private String author;private int score;private int views;
}

@SpringBootTest
@RunWith(SpringRunner.class)
public class BeanLoadServiceTest {@Autowiredprivate MongoTemplate mongoTemplate;@Testpublic void aggregateTest() {// $match阶段Criteria criteria = Criteria.where("author").is("dave");MatchOperation match = Aggregation.match(criteria);Aggregation aggregation = Aggregation.newAggregation(match);// 执行聚合管道操作AggregationResults<AggregationResult> results= mongoTemplate.aggregate(aggregation, Article.class, AggregationResult.class);List<AggregationResult> mappedResults = results.getMappedResults();// 打印结果mappedResults.forEach(System.out::println);//AggregationResult(id=512bc95fe835e68f199c8686, author=dave, score=80, views=100)//AggregationResult(id=512bc962e835e68f199c8687, author=dave, score=85, views=521)}
}

3. $match 示例

使用 $match 管道操作符选择要处理的文档，然后将结果导入到 $group 管道操作符，以计算文档的数量：

db.articles.aggregate( [// 第一阶段{ $match: { $or: [ { score: { $gt: 70, $lt: 90 } }, { views: { $gte: 1000 } } ] } },// 第二阶段{ $group: { _id: null, count: { $sum: 1 } } }
] );

第一阶段：

$match 阶段选择 score 大于 70 但小于 90 或 views 大于或等于 1000 的文档。

第二阶段：

将$match阶段筛选的文档通过管道传送到‘$group` 阶段进行计数。

// 1
{"_id": null,"count": 5
}

SpringBoot 整合 MongoDB 实现：

@Data
@Document(collection = "articles")
public class Article {@Idprivate String id;private String author;private int score;private int views;
}

@Data
public class AggregationResult {private String id;private Integer count;
}

@SpringBootTest
@RunWith(SpringRunner.class)
public class BeanLoadServiceTest {@Autowiredprivate MongoTemplate mongoTemplate;@Testpublic void aggregateTest() {// 第一阶段Criteria criteria = new Criteria();criteria.orOperator(Criteria.where("score").gt(70).lt(90), Criteria.where("views").gte(1000));MatchOperation match = Aggregation.match(criteria);// 第二阶段GroupOperation group = Aggregation.group().count().as("count");// 组合上面的2个阶段Aggregation aggregation = Aggregation.newAggregation(match,group);// 执行聚合管道操作AggregationResults<AggregationResult> results= mongoTemplate.aggregate(aggregation, Article.class, AggregationResult.class);List<AggregationResult> mappedResults = results.getMappedResults();// 打印结果mappedResults.forEach(System.out::println);// AggregationResult(id=null, count=5)}
}

2. $sort 聚合阶段

$sort 将所有输入文档进行排序，然后按照排序将其返回至管道。

{ $sort: { <field1>: <sort order>, <field2>: <sort order> ... } }

$sort 接受排序依据的字段及相应排序顺序的文档。当 sort order=1时升序排序，sort order=-1 降序排序。

如果对多个字段进行排序，则按从左到右的顺序进行排序。例如，在上面的表单中，文档首先按 field1 排序。然后，具有相同 field1 值的文档将按 field2 进一步排序。

1. 排序一致性问题

MongoDB 不按特定顺序将文档存储在集合中。对包含重复值的字段进行排序时，可能会以任何顺序返回包含这些值的文档。如果需要一致的排序顺序，请在排序中至少纳入一个包含唯一值的字段。

db.restaurants.drop()db.restaurants.insertMany( [{ "_id" : 1, "name" : "Central Park Cafe", "borough" : "Manhattan"},{ "_id" : 2, "name" : "Rock A Feller Bar and Grill", "borough" : "Queens"},{ "_id" : 3, "name" : "Empire State Pub", "borough" : "Brooklyn"},{ "_id" : 4, "name" : "Stan's Pizzaria", "borough" : "Manhattan"},{ "_id" : 5, "name" : "Jane's Deli", "borough" : "Brooklyn"},
] )

以下命令使用 $sort 阶段对 borough 字段进行排序：

db.restaurants.aggregate([{ $sort : { borough : 1 } }]
)

在此示例中，排序顺序可能不一致，因为 borough 字段包含 Manhattan 和 Brooklyn 的重复值。文档按 borough 的字母顺序返回，但具有 borough 重复值的文档的顺序在多次执行同一排序中可能不相同。

要实现一致的排序，可以在排序中添加一个仅包含唯一值的字段。以下命令使用 $sort 阶段对 borough 字段和 _id 字段进行排序：

db.restaurants.aggregate([{ $sort : { borough : 1, _id: 1 } }]
)

由于 _id 字段始终保证包含唯一值，因此在同一排序的多次执行中返回的排序顺序将始终相同。

2. $sort 示例

db.articles.drop()db.articles.insertMany([{"_id": ObjectId("512bc95fe835e68f199c8686"),"author": "dave","score": 80,"views": 100},{"_id": ObjectId("512bc962e835e68f199c8687"),"author": "dave","score": 85,"views": 521},{"_id": ObjectId("55f5a192d4bede9ac365b257"),"author": "ahn","score": 60,"views": 1000},{"_id": ObjectId("55f5a192d4bede9ac365b258"),"author": "li","score": 55,"views": 5000},{"_id": ObjectId("55f5a1d3d4bede9ac365b259"),"author": "annT","score": 55,"views": 50},{"_id": ObjectId("55f5a1d3d4bede9ac365b25a"),"author": "li","score": 94,"views": 999},{"_id": ObjectId("55f5a1d3d4bede9ac365b25b"),"author": "ty","score": 95,"views": 1000}
])

对于要作为排序依据的一个或多个字段，可以将排序顺序设置为 1 或 -1 以分别指定升序或降序。

db.articles.aggregate([{ $sort : { score : -1, views: 1 } }]
)

// 1
{"_id": ObjectId("55f5a1d3d4bede9ac365b25b"),"author": "ty","score": 95,"views": 1000
}// 2
{"_id": ObjectId("55f5a1d3d4bede9ac365b25a"),"author": "li","score": 94,"views": 999
}// 3
{"_id": ObjectId("512bc962e835e68f199c8687"),"author": "dave","score": 85,"views": 521
}// 4
{"_id": ObjectId("512bc95fe835e68f199c8686"),"author": "dave","score": 80,"views": 100
}// 5
{"_id": ObjectId("55f5a192d4bede9ac365b257"),"author": "ahn","score": 60,"views": 1000
}// 6
{"_id": ObjectId("55f5a1d3d4bede9ac365b259"),"author": "annT","score": 55,"views": 50
}// 7
{"_id": ObjectId("55f5a192d4bede9ac365b258"),"author": "li","score": 55,"views": 5000
}

SpringBoot 整合 MongoDB：

@Data
@Document(collection = "articles")
public class Article {@Idprivate String id;private String author;private int score;private int views;
}@Data
public class AggregationResult {@Idprivate String id;private String author;private int score;private int views;
}

@SpringBootTest
@RunWith(SpringRunner.class)
public class BeanLoadServiceTest {@Autowiredprivate MongoTemplate mongoTemplate;@Testpublic void aggregateTest() {// $sort阶段SortOperation sortOperation = Aggregation.sort(Sort.by(Sort.Direction.DESC, "score")).and(Sort.by(Sort.Direction.ASC, "views"));Aggregation aggregation = Aggregation.newAggregation(sortOperation);// 执行聚合查询AggregationResults<AggregationResult> results= mongoTemplate.aggregate(aggregation, Article.class, AggregationResult.class);List<AggregationResult> mappedResults = results.getMappedResults();// 打印结果mappedResults.forEach(System.out::println);//AggregationResult(id=55f5a1d3d4bede9ac365b25b, author=ty, score=95, views=1000)//AggregationResult(id=55f5a1d3d4bede9ac365b25a, author=li, score=94, views=999)//AggregationResult(id=512bc962e835e68f199c8687, author=dave, score=85, views=521)//AggregationResult(id=512bc95fe835e68f199c8686, author=dave, score=80, views=100)//AggregationResult(id=55f5a192d4bede9ac365b257, author=ahn, score=60, views=1000)//AggregationResult(id=55f5a1d3d4bede9ac365b259, author=annT, score=55, views=50)//AggregationResult(id=55f5a192d4bede9ac365b258, author=li, score=55, views=5000)}
}

3. $limit 聚合阶段

$limit 聚合阶段限制传递至管道。$limit 取一个正整数，用于指定传递的最大文档数量：

{ $limit: <positive 64-bit integer> }

如果将 $limit 阶段与以下任何一项一起使用：

• $sort 聚合阶段
• sort() 方法
• sort 命令

在将结果传递到$limit阶段之前，请务必在排序中至少包含一个包含唯一值的字段。

db.articles.drop()db.articles.insertMany([{"_id": ObjectId("512bc95fe835e68f199c8686"),"author": "dave","score": 80,"views": 100},{"_id": ObjectId("512bc962e835e68f199c8687"),"author": "dave","score": 85,"views": 521},{"_id": ObjectId("55f5a192d4bede9ac365b257"),"author": "ahn","score": 60,"views": 1000},{"_id": ObjectId("55f5a192d4bede9ac365b258"),"author": "li","score": 55,"views": 5000},{"_id": ObjectId("55f5a1d3d4bede9ac365b259"),"author": "annT","score": 60,"views": 50},{"_id": ObjectId("55f5a1d3d4bede9ac365b25a"),"author": "li","score": 94,"views": 999},{"_id": ObjectId("55f5a1d3d4bede9ac365b25b"),"author": "ty","score": 95,"views": 1000}
])

db.articles.aggregate([{ $limit : 5 }
]);

// 1
{"_id": ObjectId("512bc95fe835e68f199c8686"),"author": "dave","score": 80,"views": 100
}// 2
{"_id": ObjectId("512bc962e835e68f199c8687"),"author": "dave","score": 85,"views": 521
}// 3
{"_id": ObjectId("55f5a192d4bede9ac365b257"),"author": "ahn","score": 60,"views": 1000
}// 4
{"_id": ObjectId("55f5a192d4bede9ac365b258"),"author": "li","score": 55,"views": 5000
}// 5
{"_id": ObjectId("55f5a1d3d4bede9ac365b259"),"author": "annT","score": 55,"views": 50
}

SpringBoot 整合 MongoDB：

@Data
@Document(collection = "articles")
public class Article {@Idprivate String id;private String author;private int score;private int views;
}@Data
public class AggregationResult {@Idprivate String id;private String author;private int score;private int views;
}

@SpringBootTest
@RunWith(SpringRunner.class)
public class BeanLoadServiceTest {@Autowiredprivate MongoTemplate mongoTemplate;@Testpublic void aggregateTest() {// $sort阶段LimitOperation limitOperation = Aggregation.limit(5);Aggregation aggregation = Aggregation.newAggregation(limitOperation);// 执行聚合查询AggregationResults<AggregationResult> results= mongoTemplate.aggregate(aggregation, Article.class, AggregationResult.class);List<AggregationResult> mappedResults = results.getMappedResults();// 打印结果mappedResults.forEach(System.out::println);//AggregationResult(id=512bc95fe835e68f199c8686, author=dave, score=80, views=100)//AggregationResult(id=512bc962e835e68f199c8687, author=dave, score=85, views=521)//AggregationResult(id=55f5a192d4bede9ac365b257, author=ahn, score=60, views=1000)//AggregationResult(id=55f5a192d4bede9ac365b258, author=li, score=55, views=5000)//AggregationResult(id=55f5a1d3d4bede9ac365b259, author=annT, score=55, views=50)}
}