Mongodb 没有传统数据库的group函数，如果分组需要走MapReduce。这种MR与HadoopMR类似。下面看看Mongodb 的分组实现

现在又一张 表它的数据格式如下：
{
"_id" : ObjectId("53b224e0a1ae72328a57702c"),
"title" : "SECJ0118E",
"criteria" : "未找到对应的错误码",
"actual" : "1",
"effect" : "可能引起重大问题",
"suggestion" : "请专家提供意见",
"severity" : "Normal",
"status" : "NotOK",
"rtype" : "FormLoginExte",
"comment" : "[8/2/12 17:28:21:231 GMT+08:00] 0000001e FormLoginExte E SECJ0118E: Authentication error during authentication for user rpt",
"category" : "logs",
"time" : "0008-02-12 17:28:21"
}
{
"_id" : ObjectId("53b224e0a1ae72328a577052"),
"title" : "",
"criteria" : "未找到对应的错误码",
"actual" : "1",
"effect" : "可能引起重大问题",
"suggestion" : "请专家提供意见",
"severity" : "Normal",
"status" : "NotOK",
"rtype" : "servlet",
"comment" : "[8/2/12 19:04:41:891 CST] 0000000b servlet E com.ibm.ws.webcontainer.servlet.ServletWrapper init Uncaught.init.exception.thrown.by.servlet",
"category" : "logs",
"time" : "0008-02-12 19:04:41"
}
{
"_id" : ObjectId("53b224e0a1ae72328a576fdc"),
"title" : "系统资源设置[processes]",
"criteria" : "unlimited",
"actual" : "unlimited",
"effect" : "如果对用户资源做了limits限制，有可能造成应用运行失败或系统性能下降。",
"suggestion" : "建议修改/etc/security/limits,编辑root相关参数部分都为-1。",
"severity" : "None",
"status" : "OK",
"rtype" : "系统参数设置检查",
"comment" : "",
"category" : "params"
}
1：单个条件分组

(1) 现在我们根据title进行分组 并且统计每个组的数量

db.runCommand({ mapreduce: "check_result",
map : function Map() {
//emit 函数中的key是唯一的，是分组条件 这里把title传递过来。
emit(
this.title,
{count: 1}
);
},
reduce : function Reduce(key, values) {
total=0;//定义一个变量total , values是一个数组
for( var i in values){
total +=values[i].count
}

return {"count":total};
},
finalize : function Finalize(key, reduced) {
return reduced;
},
out : { inline : 1 }
});

结果如下（结果中的_id键就是要分组的title 。value是分组后的值）：
{ "_id" : "" , "value" : { "count" : 113.0}}
{ "_id" : "/tmp是否设置了t标志位" , "value" : { "count" : 21.0}}
{ "_id" : "ASYN0080W" , "value" : { "count" : 120.0}}
{ "_id" : "AppServer的JVM堆最大值" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的JVM堆最小值" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的JVM标准输出日志切换周期" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的JVM标准输出日志回滚类型" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的JVM标准错误日志切换周期" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的JVM标准错误日志回滚类型" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的WebContainer线程池最大值" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的WebContainer线程池最小值" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的通用JVM参数" , "value" : { "count" : 6.0}}
{ "_id" : "AppServer的通用JVM参数-SystemGC" , "value" : { "count" : 6.0}}
{ "_id" : "Audit是否打开" , "value" : { "count" : 21.0}}
{ "_id" : "CWPKI0041W" , "value" : { "count" : 65.0}}
{ "_id" : "CWPMC0017W" , "value" : { "count" : 7.0}}
{ "_id" : "CWSAA0037W" , "value" : { "count" : 13.0}}
{ "_id" : "Could not invoke an operation on object" , "value" : { "count" : 21.0}}
{ "_id" : "DCSV0000W" , "value" : { "count" : 4.0}}
{ "_id" : "DCSV1115W" , "value" : { "count" : 137.0}}

2：多个条件分组

(1) 现在我们根据title,status,severity进行分组 并且统计每个组的数量
db.runCommand({ mapreduce: "check_result",
map : function Map() {
//emit 函数中的key是唯一的，是分组条件
emit(
{"title":this.title,"status":this.status,"serverity":this.severity}
,
{count: 1}
);
},
reduce : function Reduce(key, values) {
total=0;//定义一个变量total , values是一个数组
for( var i in values){
total +=values[i].count
}

return {"count":total};
},
finalize : function Finalize(key, reduced) {
return reduced;
},
out : { inline : 1 }
});

输出结果如下格式化：

{ "_id" : { "title" : "" , "status" : "NotOK"} , "value" : { "count" : 113.0}}
{ "_id" : { "title" : "/tmp是否设置了t标志位" , "status" : "NotOK"} , "value" : { "count" : 21.0}}
{ "_id" : { "title" : "ASYN0080W" , "status" : "NotOK"} , "value" : { "count" : 120.0}}
{ "_id" : { "title" : "AppServer的JVM堆最大值" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的JVM堆最小值" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的JVM标准输出日志切换周期" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的JVM标准输出日志回滚类型" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的JVM标准错误日志切换周期" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的JVM标准错误日志回滚类型" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的WebContainer线程池最大值" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的WebContainer线程池最小值" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的通用JVM参数" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "AppServer的通用JVM参数-SystemGC" , "status" : "NotOK"} , "value" : { "count" : 6.0}}
{ "_id" : { "title" : "Audit是否打开" , "status" : "NotOK"} , "value" : { "count" : 21.0}}
{ "_id" : { "title" : "CWPKI0041W" , "status" : "NotOK"} , "value" : { "count" : 65.0}}
{ "_id" : { "title" : "CWPMC0017W" , "status" : "NotOK"} , "value" : { "count" : 7.0}}
{ "_id" : { "title" : "CWSAA0037W" , "status" : "NotOK"} , "value" : { "count" : 13.0}}
{ "_id" : { "title" : "Could not invoke an operation on object" , "status" : "NotOK"} , "value" : { "count" : 21.0}}
{ "_id" : { "title" : "DCSV0000W" , "status" : "NotOK"} , "value" : { "count" : 4.0}}
{ "_id" : { "title" : "DCSV1115W" , "status" : "NotOK"} , "value" : { "count" : 137.0}}

本文转自94cool博客园博客，原文链接：http://www.cnblogs.com/94cool/p/6435481.html，如需转载请自行联系原作者