Mongodb查询执行时间

Posted 2023-04-19

【中文标题】Mongodb查询执行时间

【英文标题】：Mongodb query execution time

【发布时间】：2019-04-27 11:58:00

【问题描述】：

我在下面使用了 Query 花费了很多时间

查询

db.saleOrder.find("currentStatus._id":"147","_id":1).limit(10).explain("executionStats")

ExecutionStats 结果

    "queryPlanner" : 
        "plannerVersion" : 1,
        "namespace" : "db_erp_tube.saleOrder",
        "indexFilterSet" : false,
        "parsedQuery" : 
            "currentStatus._id" : 
                "$eq" : "147"
            
        ,
        "winningPlan" : 
            "stage" : "LIMIT",
            "limitAmount" : 10,
            "inputStage" : 
                "stage" : "PROJECTION",
                "transformBy" : 
                    "_id" : 1
                ,
                "inputStage" : 
                    "stage" : "COLLSCAN",
                    "filter" : 
                        "currentStatus._id" : 
                            "$eq" : "147"
                        
                    ,
                    "direction" : "forward"
                
            
        ,
        "rejectedPlans" : [ ]
    ,
    "executionStats" : 
        "executionSuccess" : true,
        "nReturned" : 10,
        "executionTimeMillis" : 8673,
        "totalKeysExamined" : 0,
        "totalDocsExamined" : 3458482,
        "executionStages" : 
            "stage" : "LIMIT",
            "nReturned" : 10,
            "executionTimeMillisEstimate" : 8460,
            "works" : 3458484,
            "advanced" : 10,
            "needTime" : 3458473,
            "needYield" : 0,
            "saveState" : 27019,
            "restoreState" : 27019,
            "isEOF" : 1,
            "invalidates" : 0,
            "limitAmount" : 10,
            "inputStage" : 
                "stage" : "PROJECTION",
                "nReturned" : 10,
                "executionTimeMillisEstimate" : 8450,
                "works" : 3458483,
                "advanced" : 10,
                "needTime" : 3458473,
                "needYield" : 0,
                "saveState" : 27019,
                "restoreState" : 27019,
                "isEOF" : 0,
                "invalidates" : 0,
                "transformBy" : 
                    "_id" : 1
                ,
                "inputStage" : 
                    "stage" : "COLLSCAN",
                    "filter" : 
                        "currentStatus._id" : 
                            "$eq" : "147"
                        
                    ,
                    "nReturned" : 10,
                    "executionTimeMillisEstimate" : 8400,
                    "works" : 3458483,
                    "advanced" : 10,
                    "needTime" : 3458473,
                    "needYield" : 0,
                    "saveState" : 27019,
                    "restoreState" : 27019,
                    "isEOF" : 0,
                    "invalidates" : 0,
                    "direction" : "forward",
                    "docsExamined" : 3458482
                
            
        
    ,
    "serverInfo" : 
        "host" : "172.16.109",
        "port" : 27017,
        "version" : "4.0.0",
        "gitVersion" : "3b07af3d4f471ae89e8186d33bbb1d5259597d51"
    ,
    "ok" : 1,
    "operationTime" : Timestamp(1556365275, 114),
    "$clusterTime" : 
        "clusterTime" : Timestamp(1556365275, 114),
        "signature" : 
            "hash" : BinData(0,"ppu91nKmeiC//+UvdsEbjrBTDLU="),
            "keyId" : NumberLong("6633468944474701825")
        
    

【问题讨论】：

尝试索引 currentStatus._id 然后检查 stat！

【参考方案1】：

该查询的执行时间超过 8 秒（8673 毫秒），因为它必须在返回 10 个文档之前扫描 saleOrder 集合中的所有 3458482 个文档。

这在解释输出的过滤阶段进行了说明。

"inputStage" : 
                "stage" : "COLLSCAN",
                "filter" : 
                    "currentStatus._id" : 
                        "$eq" : "147"
                    
                ,

COLLSCAN 表示全集合扫描，

 "totalDocsExamined" : 3458482,

这是在返回结果之前扫描的文档数。

 "executionTimeMillis" : 8673,

这是运行查询所用的总时间。

正如@the_mahasagar 建议的那样，在 currentStatus._id 上建立索引可以大大加快查询速度，请使用以下命令。

 db.saleOrder.createIndex( "currentStatus._id": 1)