MongoDB的使用学习之(七)MongoDB的聚合查询(两种方式)附项目源码
先来张在路上……
铛铛铛……项目源码下载地址:https://files.cnblogs.com/ontheroad_lee/MongoDBDemo.rar
此项目是用Maven创建的,没有使用Maven的,自己百度、谷歌去;直接用Junit测试就行,先执行里面的save方法,添加10000条测试数据提供各种聚合查询等。
少废话,上干货……
一、MongoDB数据库的配置(mongodb.xml)
以下是我自己的配置,红色字体请改为自己本机的东东,你说不懂设置端口,不会创建数据库名称,不会配置用户名密码,那有请查阅本系列的第4节(MongoDB的使用学习之(四)权限设置--用户名、密码、端口==),你说懒得设置,那就@#¥%……&*()!
<!-- Default bean name is 'mongo' --> <!-- 定义mongo对象,对应的是mongodb官方jar包中的Mongo,replica-set设置集群副本的ip地址和端口 --> <mongo:mongo id="mongo" host="localhost" port="47017" /> <mongo:db-factory id="mongoDbFactory" dbname="mongoTest" mongo-ref="mongo" username="root1" password="root1" /> <!-- 自动扫描以下包的类 --> <!-- 映射转换器,扫描back-package目录下的文件,根据注释,把它们作为mongodb的一个collection的映射 --> <mongo:mapping-converter base-package="com.ontheroad.**.po" /> <bean id="mappingContext" class="org.springframework.data.mongodb.core.mapping.MongoMappingContext" /> <!-- 配置mongodb映射类型 --> <bean id="mappingMongoConverter" class="org.springframework.data.mongodb.core.convert.MappingMongoConverter"> <constructor-arg name="mongoDbFactory" ref="mongoDbFactory" /> <constructor-arg name="mappingContext" ref="mappingContext" /> <property name="typeMapper" ref="defaultMongoTypeMapper" /> </bean> <!-- 默认Mongodb类型映射 --> <bean id="defaultMongoTypeMapper" class="org.springframework.data.mongodb.core.convert.DefaultMongoTypeMapper"> <constructor-arg name="typeKey"> <null /><!-- 这里设置为空,可以把 spring data mongodb 多余保存的_class字段去掉 --> </constructor-arg> </bean> <!-- 操作mongodb --> <!-- mongodb的主要操作对象mongoTemplate(Spring Data提供),所有对mongodb的增删改查的操作都是通过它完成 --> <bean id="mongoTemplate" class="org.springframework.data.mongodb.core.MongoTemplate"> <constructor-arg name="mongoDbFactory" ref="mongoDbFactory" /> <constructor-arg name="mongoConverter" ref="mappingMongoConverter" /> </bean>
二、各种聚合查询方法
以下只是展示一些常用的,聚合查询方法,无奈个人功力尚浅,没啥高深的东西,待日后,有时间有精力有实力,再整理些高级一点的
1、添加测试数据
@Test public void save() { News n = null; for (int i = 0; i < 10000; i++) { n = new News(); n.setTitle("title_" + i); n.setUrl("url_" + i); //2014-01-01到2014-01-01之间的随机时间 Date randomDate=DateUtil.randomDate("2014-01-01","2014-05-11"); //MongoDB里如果时间类型存的是Date,那么会差8个小时的时区,因为MongoDB使用的格林威治时间,中国所处的是+8区,so…… //比如我保存的是2014-05-01 00:00:00,那么保存到MongoDB里则是2014-05-01 08:00:00,所以为了统一方面,那就保存字符串类型,底下保存的long类型 n.setPublishTimeStr(DateUtil.formatDateTimeByDate(randomDate)); //long类型在查询速度中肯定会比较快 n.setPublishTime(randomDate.getTime()); n.setPublishDate(randomDate); n.setPublishMedia("publishMedia_" + i); String[] areaArr = {"1024", "102401", "102402", "102403", "102404", "102405", "102406", "102407", "102408" , "10240101", "10240102", "10240201", "10240202", "10240301", "10240302", "10240401", "10240402" , "10240501", "10240502", "10240601", "10240602", "10240701", "10240702", "10240801", "10240802"}; int areaNum=(int)(Math.random() * areaArr.length);//产生0-strs.length的整数随机数 String area = areaArr[areaNum]; n.setArea(area); String[] ckeyArr = {"A101", "A102", "A201", "A202", "A203" , "B101", "B102", "B103", "C201", "C202", "C203", "22", "23", "24", "25", "26"}; int ckeyNum=(int)(Math.random() * ckeyArr.length);//产生0-strs.length的整数随机数 List<String> list = new ArrayList<String>(); for (int j = 0; j < ckeyNum; j ++) { int ckeyNum1=(int)(Math.random() * ckeyArr.length);//产生0-strs.length的整数随机数 list.add(ckeyArr[ckeyNum1]); } n.setClassKey(list); Integer[] evalArr = {1, 0}; int evalNum=(int)(Math.random() * evalArr.length);//产生0-strs.length的整数随机数 n.setEvaluate(evalArr[evalNum]); Integer[] mproArr = {1, 2, 100}; int mproNum=(int)(Math.random() * mproArr.length);//产生0-strs.length的整数随机数 n.setMediaProperty(mproArr[mproNum]); Integer[] mtypeArr = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11}; int mtypeNum=(int)(Math.random() * mtypeArr.length);//产生0-strs.length的整数随机数 n.setMediaType(mtypeArr[mtypeNum]); Integer[] levelArr = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12}; int levelNum=(int)(Math.random() * levelArr.length);//产生0-strs.length的整数随机数 n.setLevel(levelArr[levelNum]); newsService.save(n); } System.out.println("OK"); }
2、简单的分组查询--使用Mongo本身提供的AggregationOutput进行分组查询
/** * * 功能:使用Mongo本身提供的AggregationOutput进行分组查询 * 参数: * 创建人:OnTheRoad_Lee * 修改人:OnTheRoad_Lee * 最后修改时间:2014-5-26 */ public void testGroup1 () { //按照eval字段进行分组,注意$eval必须是存在mongodb里面的字段,不能写$evaluate(此字段是News类中定义的,和存入mongodb中的有区别) //{$group:{_id:{'AAA':'$BBB'},CCC:{$sum:1}}}固定格式:把要分组的字段放在_id:{}里面,BBB是mongodb里面的某个字段,AAA是BBB的重命名,CCC是$sum:1的重命名 //此查询语句== select eval as eval, count(*) as docsNum from news group by eval having docsNum>=85 order by docsNum desc //具体的mongodb和sql的对照可以参考:http://docs.mongodb.org/manual/reference/sql-aggregation-comparison/ String groupStr = "{$group:{_id:{'eval':'$eval'},docsNum:{$sum:1}}}"; DBObject group = (DBObject) JSON.parse(groupStr); String matchStr = "{$match:{docsNum:{$gte:85}}}"; DBObject match = (DBObject) JSON.parse(matchStr); String sortStr = "{$sort:{_id.docsNum:-1}}"; DBObject sort = (DBObject) JSON.parse(sortStr); AggregationOutput output = mongoTemplate.getCollection("news").aggregate(group, match, sort); System.out.println(output.getCommand()); //转换为执行原生的mongodb查询语句 //{ "aggregate" : "news" , "pipeline" : [ { "$group" : { "_id" : { "eval" : "$eval"} , "docsNum" : { "$sum" : 1}}} , { "$match" : { "docsNum" : { "$gte" : 85}}} , { "$sort" : { "_id.docsNum" : -1}}]} System.out.println(output.getCommandResult()); //查询结果 //{ "serverUsed" : "localhost/127.0.0.1:47017" , "result" : [ { "_id" : { "evaluate" : 1} , "docsNum" : 9955} , { "_id" : { "evaluate" : 0} , "docsNum" : 10047}] , "ok" : 1.0} //也可以把查询结果封装到NewsNumDTO,这样以一个dto对象返回前台操作就更容易了 NewsNumDTO dto = new NewsNumDTO(); for( Iterator< DBObject > it = output.results().iterator(); it.hasNext(); ){ BasicDBObject dbo = ( BasicDBObject ) it.next(); BasicDBObject keyValus = (BasicDBObject)dbo.get("_id"); int eval = keyValus.getInt("eval"); long docsNum = ((Integer)dbo.get("docsNum")).longValue(); if(eval == 1){ dto.setPositiveNum(docsNum); }else { dto.setNegativeNum(docsNum); } } }
3、获取和testGroup1方法同样结果的另一种写法,Spring Data MongoDB隆重登场,语法更加简洁易懂
/** * * 功能:获取和testGroup1方法同样结果的另一种写法,Spring Data MongoDB隆重登场,语法更加简洁易懂 * 参数: * 创建人:OnTheRoad_Lee * 修改人:OnTheRoad_Lee * 最后修改时间:2014-5-26 */ public void testAggregation1() { TypedAggregation<News> agg = Aggregation.newAggregation( News.class, project("evaluate") ,group("evaluate").count().as("totalNum") ,match(Criteria.where("totalNum").gte(85)) ,sort(Sort.Direction.DESC, "totalNum") ); AggregationResults<BasicDBObject> result = mongoTemplate.aggregate(agg, BasicDBObject.class); System.out.println(agg.toString()); //执行语句差不多 //{ "aggregate" : "__collection__" , "pipeline" : [ { "$project" : { "evaluate" : 1}} , { "$group" : { "_id" : "$evaluate" , "totalNum" : { "$sum" : 1}}} , { "$match" : { "totalNum" : { "$gte" : 85}}} , { "$sort" : { "totalNum" : -1}}]} System.out.println(result.getMappedResults()); //查询结果简洁明了 //[{ "_id" : 0 , "totalNum" : 10047}, { "_id" : 1 , "totalNum" : 9955}] //使用此方法,如果封装好了某一个类,类里面的属性和结果集的属性一一对应,那么,Spring是可以直接把结果集给封装进去的 //就是AggregationResults<BasicDBObject> result = mongoTemplate.aggregate(agg, BasicDBObject);中的BasicDBObject改为自己封装的类 //但是感觉这样做有点不灵活,其实吧,应该是自己现在火候还不到,还看不到他的灵活性,好处在哪里;等火候旺了再说呗 //所以,就用这个万能的BasicDBObject类来封装返回结果 List<BasicDBObject> resultList = result.getMappedResults(); NewsNumDTO dto = new NewsNumDTO(); for(BasicDBObject dbo : resultList){ int eval = dbo.getInt("_id"); long num = dbo.getLong("totalNum"); if(eval == 1){ dto.setPositiveNum(num); }else { dto.setNegativeNum(num); } } System.out.println(dto.getPositiveNum()); }
4、previousOperation的简单使用--为分组的字段(_id)建立别名
/** * * 功能:previousOperation的简单使用--为分组的字段(_id)建立别名 * 参数: * 创建人:OnTheRoad_Lee * 修改人:OnTheRoad_Lee * 最后修改时间:2014-5-26 */ public void testAggregation2() { TypedAggregation<News> agg = Aggregation.newAggregation( News.class, // match(Criteria.where("mediaType").is(100)) project("evaluate") ,group("evaluate").count().as("totalNum") ,project("evaluate", "totalNum") .and("eval").previousOperation() //为分组的字段(_id)建立别名 ); AggregationResults<BasicDBObject> result = mongoTemplate.aggregate(agg, BasicDBObject.class); System.out.println(agg.toString()); // { "aggregate" : "__collection__" , "pipeline" : [ { "$project" : { "evaluate" : 1}} , { "$group" : { "_id" : "$evaluate" , "totalNum" : { "$sum" : 1}}} , { "$project" : { "evaluate" : "$_id.evaluate" , "totalNum" : 1 , "_id" : 0 , "eval" : "$_id"}}]} System.out.println(result.getMappedResults()); // [{ "totalNum" : 10047 , "eval" : 0}, { "totalNum" : 9955 , "eval" : 1}] }
5、unwind()的使用,通过Spring Data MongoDB
/** * * 功能:unwind()的使用,通过Spring Data MongoDB * unwind()就是$unwind这个命令的转换, * $unwind - 可以将一个包含数组的文档切分成多个, 比如你的文档有 中有个数组字段 A, A中有10个元素, 那么 * 经过 $unwind处理后会产生10个文档,这些文档只有 字段 A不同 * 详见:http://my.oschina.net/GivingOnenessDestiny/blog/88006 * 参数: * 创建人:OnTheRoad_Lee * 修改人:OnTheRoad_Lee * 最后修改时间:2014-5-26 */ public void testAggregation3() { TypedAggregation<News> agg = Aggregation.newAggregation( News.class, unwind("classKey") ,project("evaluate", "classKey") // 这里说明一点就是如果group>=2个字段,那么结果集的分组字段就没有_id了,取而代之的是具体的字段名(和testAggregation()最对比) ,group("evaluate", "classKey").count().as("totalNum") ,sort(Sort.Direction.DESC, "totalNum") ); AggregationResults<BasicDBObject> result = mongoTemplate.aggregate(agg, BasicDBObject.class); System.out.println(agg.toString()); // { "aggregate" : "__collection__" , "pipeline" : [ { "$unwind" : "$classKey"} , { "$project" : { "evaluate" : 1 , "classKey" : 1}} , { "$group" : { "_id" : { "evaluate" : "$evaluate" , "classKey" : "$classKey"} , "totalNum" : { "$sum" : 1}}} , { "$sort" : { "totalNum" : -1}}]} System.out.println(result.getMappedResults()); // 结果就是酱紫,一目了然,怎么操作,就交给你自己了 // [{ "evaluate" : 0 , "classKey" : "A201" , "totalNum" : 4857}, { "evaluate" : 0 , "classKey" : "23" , "totalNum" : 4857}, { "evaluate" : 0 , "classKey" : "A101" , "totalNum" : 4842}, { "evaluate" : 0 , "classKey" : "24" , "totalNum" : 4806}, { "evaluate" : 1 , "classKey" : "A101" , "totalNum" : 4787}, { "evaluate" : 0 , "classKey" : "C201" , "totalNum" : 4787}, { "evaluate" : 0 , "classKey" : "A102" , "totalNum" : 4783},……] }
6、Spring Data MongoDB,按照时间(字符串)分组
/** * * 功能:Spring Data MongoDB,按照时间(字符串)分组 * 参数: * 创建人:OnTheRoad_Lee * 修改人:OnTheRoad_Lee * 最后修改时间:2014-5-26 */ public void testAggregation4() { TypedAggregation<News> agg = Aggregation.newAggregation( News.class, //project().andExpression()里面是一个表达式 // 详见api:http://docs.spring.io/spring-data/data-mongodb/docs/current/reference/htmlsingle/#mongo.aggregation // 搜索 .andExpression 定位到具体的方法模块 project("evaluate") .andExpression("substr(publishTimeStr,0,10)").as("publishDate") ,group("evaluate", "publishDate").count().as("totalNum") ,sort(Sort.Direction.DESC, "totalNum") ); AggregationResults<BasicDBObject> result = mongoTemplate.aggregate(agg, BasicDBObject.class); System.out.println(agg.toString()); // { "aggregate" : "__collection__" , "pipeline" : [ { "$project" : { "evaluate" : 1 , "publishDate" : { "$substr" : [ "$publishTimeStr" , 0 , 10]}}} , { "$group" : { "_id" : { "evaluate" : "$evaluate" , "publishDate" : "$publishDate"} , "totalNum" : { "$sum" : 1}}} , { "$sort" : { "totalNum" : -1}}]} System.out.println(result.getMappedResults()); // [{ "evaluate" : 0 , "publishDate" : "2014-03-09" , "totalNum" : 101}, { "evaluate" : 1 , "publishDate" : "2014-02-14" , "totalNum" : 100}, { "evaluate" : 1 , "publishDate" : "2014-02-11" , "totalNum" : 99}, { "evaluate" : 0 , "publishDate" : "2014-03-17" , "totalNum" : 98}, { "evaluate" : 1 , "publishDate" : "2014-03-26" , "totalNum" : 98}, ……] // 这个查询结果貌似不是我们想要的效果,理想,一目了然的效果应该是,以日期为单位,日期底下正面多少,负面多少: // [ // { "publishDate" : "2014-03-09" , "evalInfo" : [{"evaluate" : 0 , "totalNum" : 101}, {"evaluate" : 1 , "totalNum" : 44}]} // { "publishDate" : "2014-03-12" , "evalInfo" : [{"evaluate" : 0 , "totalNum" : 11}, {"evaluate" : 1 , "totalNum" : 32}]}, // …… // ] // 无奈本人功力尚浅,查了N多资料,各种论坛,中文的,英文的都查了,就是找不到Spring Data MongoDB 分组方法 // ,所以就引出了testAggregation5 }
7、使用原生态mongodb语法,按照时间(字符串)分组,整合查询结果,使用$push命令
/** * * 功能:使用原生态mongodb语法,按照时间(字符串)分组,整合查询结果,使用$push命令 * 参数: * 创建人:OnTheRoad_Lee * 修改人:OnTheRoad_Lee * 最后修改时间:2014-5-26 */ public void testAggregation5() { /* Group操作*/ String groupStr = "{$group:{_id:{'evaluate':'$eval','ptimes':{$substr:['$ptimes',0,10]}},totalNum:{$sum:1}}}"; DBObject group = (DBObject) JSON.parse(groupStr); /* Reshape Group Result*/ DBObject projectFields = new BasicDBObject(); projectFields.put("ptimes", "$_id.ptimes"); projectFields.put("evalInfo", new BasicDBObject("evaluate","$_id.evaluate").append("totalNum", "$totalNum")); DBObject project = new BasicDBObject("$project", projectFields); /* 将结果push到一起*/ DBObject groupAgainFields = new BasicDBObject("_id", "$ptimes"); groupAgainFields.put("evalInfo", new BasicDBObject("$push", "$evalInfo")); DBObject reshapeGroup = new BasicDBObject("$group", groupAgainFields); /* 查看Group结果 */ AggregationOutput output = mongoTemplate.getCollection("news").aggregate(group, project, reshapeGroup); System.out.println(output.getCommand()); // { "aggregate" : "news" , "pipeline" : [ { "$group" : { "_id" : { "evaluate" : "$eval" , "ptimes" : { "$substr" : [ "$ptimes" , 0 , 10]}} , "totalNum" : { "$sum" : 1}}} , { "$project" : { "ptimes" : "$_id.ptimes" , "evalInfo" : { "evaluate" : "$_id.evaluate" , "totalNum" : "$totalNum"}}} , { "$group" : { "_id" : "$ptimes" , "evalInfo" : { "$push" : "$evalInfo"}}}]} System.out.println(output.getCommandResult()); // { "serverUsed" : "localhost/127.0.0.1:47017" , "result" : [ // { "_id" : "2014-04-24" , "evalInfo" : [ { "evaluate" : 1 , "totalNum" : 67} , { "evaluate" : 0 , "totalNum" : 76}]} // , { "_id" : "2014-02-05" , "evalInfo" : [ { "evaluate" : 1 , "totalNum" : 70} , { "evaluate" : 0 , "totalNum" : 84}]} // , { "_id" : "2014-03-21" , "evalInfo" : [ { "evaluate" : 0 , "totalNum" : 82} , { "evaluate" : 1 , "totalNum" : 89}]}}……] // , "ok" : 1.0} }