在MongoDB中搜索两个字段比搜索一个字段要慢得多_Mongodb_Performance_Indexing

在MongoDB中搜索两个字段比搜索一个字段要慢得多

mongodb performance indexing

在MongoDB中搜索两个字段比搜索一个字段要慢得多,mongodb,performance,indexing,Mongodb,Performance,Indexing,我是MongoDB的新手，尝试用相当大的数据库设置1个节点。10亿行，140GB，10个字段。除一个字段外，所有字段都有单独的索引，因为我们计划了许多不同的查询，这些查询将组合2或3个字段我的问题是MongoDB似乎没有在组合搜索中使用单独的索引 db.alldata.find({ p: "степанович" }) 此搜索的回答时间不到1秒 db.alldata.find({ n: "иван" }) 这个也是但是db.alldata.find（{n:“ааа”，p:“ааааааа

我是MongoDB的新手，尝试用相当大的数据库设置1个节点。10亿行，140GB，10个字段。除一个字段外，所有字段都有单独的索引，因为我们计划了许多不同的查询，这些查询将组合2或3个字段

我的问题是MongoDB似乎没有在组合搜索中使用单独的索引

db.alldata.find({ p: "степанович" })

此搜索的回答时间不到1秒

db.alldata.find({ n: "иван" })

这个也是

但是db.alldata.find（{n:“ааа”，p:“ааааааааач”}）

要花很长时间才能完成。我做错什么了吗？我不想使用组合索引，因为我们有太多不同的查询，甚至单独的索引几乎占用所有RAM（50GB）

将创建索引。而且有足够的内存来容纳它们。 OS是CentOS7，FS是xfs（如果有必要的话）
更新：

db.alldata.find({ n: "иван", p: "степанович"}).explain("allPlansExecution") { "queryPlanner" : { "plannerVersion" : 1, "namespace" : "myDb.alldata", "indexFilterSet" : false, "parsedQuery" : { "$and" : [ { "n" : { "$eq" : "иван" } }, { "p" : { "$eq" : "степанович" } } ] }, "winningPlan" : { "stage" : "FETCH", "filter" : { "n" : { "$eq" : "иван" } }, "inputStage" : { "stage" : "IXSCAN", "keyPattern" : { "p" : 1 }, "indexName" : "parentalName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "p" : [ "[\"степанович\", \"степанович\"]" ] } } }, "rejectedPlans" : [ { "stage" : "FETCH", "filter" : { "p" : { "$eq" : "степанович" } }, "inputStage" : { "stage" : "IXSCAN", "keyPattern" : { "n" : 1 }, "indexName" : "firstName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "n" : [ "[\"иван\", \"иван\"]" ] } } }, { "stage" : "FETCH", "filter" : { "$and" : [ { "n" : { "$eq" : "иван" } }, { "p" : { "$eq" : "степанович" } } ] }, "inputStage" : { "stage" : "AND_SORTED", "inputStages" : [ { "stage" : "IXSCAN", "keyPattern" : { "n" : 1 }, "indexName" : "firstName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "n" : [ "[\"иван\", \"иван\"]" ] } }, { "stage" : "IXSCAN", "keyPattern" : { "p" : 1 }, "indexName" : "parentalName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "p" : [ "[\"степанович\", \"степанович\"]" ] } } ] } } ] }, "executionStats" : { "executionSuccess" : true, "nReturned" : 31, "executionTimeMillis" : 213504, "totalKeysExamined" : 124244, "totalDocsExamined" : 124244, "executionStages" : { "stage" : "FETCH", "filter" : { "n" : { "$eq" : "иван" } }, "nReturned" : 31, "executionTimeMillisEstimate" : 13160, "works" : 124246, "advanced" : 31, "needTime" : 124213, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 1, "invalidates" : 0, "docsExamined" : 124244, "alreadyHasObj" : 0, "inputStage" : { "stage" : "IXSCAN", "nReturned" : 124244, "executionTimeMillisEstimate" : 240, "works" : 124245, "advanced" : 124244, "needTime" : 0, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 1, "invalidates" : 0, "keyPattern" : { "p" : 1 }, "indexName" : "parentalName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "p" : [ "[\"степанович\", \"степанович\"]" ] }, "keysExamined" : 124244, "dupsTested" : 0, "dupsDropped" : 0, "seenInvalidated" : 0 } }, "allPlansExecution" : [ { "nReturned" : 18, "executionTimeMillisEstimate" : 196410, "totalKeysExamined" : 124245, "totalDocsExamined" : 124245, "executionStages" : { "stage" : "FETCH", "filter" : { "p" : { "$eq" : "степанович" } }, "nReturned" : 18, "executionTimeMillisEstimate" : 196410, "works" : 124245, "advanced" : 18, "needTime" : 124227, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 0, "invalidates" : 0, "docsExamined" : 124245, "alreadyHasObj" : 0, "inputStage" : { "stage" : "IXSCAN", "nReturned" : 124245, "executionTimeMillisEstimate" : 240, "works" : 124245, "advanced" : 124245, "needTime" : 0, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 0, "invalidates" : 0, "keyPattern" : { "n" : 1 }, "indexName" : "firstName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "n" : [ "[\"иван\", \"иван\"]" ] }, "keysExamined" : 124245, "dupsTested" : 0, "dupsDropped" : 0, "seenInvalidated" : 0 } } }, { "nReturned" : 13, "executionTimeMillisEstimate" : 540, "totalKeysExamined" : 124245, "totalDocsExamined" : 13, "executionStages" : { "stage" : "FETCH", "filter" : { "$and" : [ { "n" : { "$eq" : "иван" } }, { "p" : { "$eq" : "степанович" } } ] }, "nReturned" : 13, "executionTimeMillisEstimate" : 540, "works" : 124245, "advanced" : 13, "needTime" : 124232, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 0, "invalidates" : 0, "docsExamined" : 13, "alreadyHasObj" : 0, "inputStage" : { "stage" : "AND_SORTED", "nReturned" : 13, "executionTimeMillisEstimate" : 490, "works" : 124245, "advanced" : 13, "needTime" : 124232, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 0, "invalidates" : 0, "flagged" : 0, "failedAnd_0" : 29865, "failedAnd_1" : 29858, "inputStages" : [ { "stage" : "IXSCAN", "nReturned" : 74468, "executionTimeMillisEstimate" : 250, "works" : 74468, "advanced" : 74468, "needTime" : 0, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 0, "invalidates" : 0, "keyPattern" : { "n" : 1 }, "indexName" : "firstName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "n" : [ "[\"иван\", \"иван\"]" ] }, "keysExamined" : 74468, "dupsTested" : 0, "dupsDropped" : 0, "seenInvalidated" : 0 }, { "stage" : "IXSCAN", "nReturned" : 49777, "executionTimeMillisEstimate" : 150, "works" : 49777, "advanced" : 49777, "needTime" : 0, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 0, "invalidates" : 0, "keyPattern" : { "p" : 1 }, "indexName" : "parentalName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "p" : [ "[\"степанович\", \"степанович\"]" ] }, "keysExamined" : 49777, "dupsTested" : 0, "dupsDropped" : 0, "seenInvalidated" : 0 } ] } } }, { "nReturned" : 31, "executionTimeMillisEstimate" : 13160, "totalKeysExamined" : 124244, "totalDocsExamined" : 124244, "executionStages" : { "stage" : "FETCH", "filter" : { "n" : { "$eq" : "иван" } }, "nReturned" : 31, "executionTimeMillisEstimate" : 13160, "works" : 124245, "advanced" : 31, "needTime" : 124213, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 1, "invalidates" : 0, "docsExamined" : 124244, "alreadyHasObj" : 0, "inputStage" : { "stage" : "IXSCAN", "nReturned" : 124244, "executionTimeMillisEstimate" : 240, "works" : 124245, "advanced" : 124244, "needTime" : 0, "needYield" : 0, "saveState" : 11965, "restoreState" : 11965, "isEOF" : 1, "invalidates" : 0, "keyPattern" : { "p" : 1 }, "indexName" : "parentalName", "isMultiKey" : false, "isUnique" : false, "isSparse" : false, "isPartial" : false, "indexVersion" : 1, "direction" : "forward", "indexBounds" : { "p" : [ "[\"степанович\", \"степанович\"]" ] }, "keysExamined" : 124244, "dupsTested" : 0, "dupsDropped" : 0, "seenInvalidated" : 0 } } } ] }, "serverInfo" : { "host" : "cass", "port" : 27017, "version" : "3.2.4", "gitVersion" : "e2ee9ffcf9f5a94fad76802e28cc978718bb7a30" }, "ok" : 1 }
我的问题是MongoDB似乎没有在上使用单独的索引联合搜索
在MongoDb 2.6之后，MongoDb支持
但如果需要，应该使用联合索引。因为它可以很好地处理
即使使用索引，查询也非常重要。因此，您应该使用基于的查询。因此，您应该选择将使用哪个索引，或者不使用索引进行查询

理解这些概念后，您可以根据自己的期望做出最佳选择。
请更新您的帖子，并解释结果<代码>db.alldata.find（{n:“аааа”，p:“ааааааааааач”}）。另外，如果您使用的是mmap或wiredtiger引擎，请说明您的MongoDB版本。官方回购和wiredtiger的“3.2.4”。添加了所有计划执行。谢谢。请将
collection.getIndexes（）
的输出也发送给我。制作复合索引是浪费时间。没有任何变化，3个字段的复合索引为32GB，而3个单独的索引只占17。现在重新构建单独的索引（前台大约需要10个小时），并将使用提示进行实验。问题中描述了DB。10个字段，每个字段占用7-10GBs，所以我买不起复合索引。他们不适合RAM。难以置信，但这没有帮助。我做db.alldata.find（blabla.hint）（“姓氏”，“名字”）和查询工作几分钟。当然，在我的10亿行中，我有许多非唯一的值。我无法改变这一点。很有趣。MongoDB允许按名称指定带有两个索引的提示，但实际上不使用第二个索引。并且不允许按列名在提示中设置第二个索引。所以，我可能也必须使用复合索引。你的moongodb版本是什么？“版本”：“3.2.4”，“gitVersion”：“e2ee9ffcf9f5a94fad76802e28cc978718bb7a30”