mongodb - mongodb 4.2 按 id 排序慢，复合索引
问题描述
我有时尚的文档
{
_id: ...,
p: [
{
k: 'stringvalue',
v: 'stringvalue'
},
...
]
}
_id提供给驱动程序作为插入的序列号，例如...0001，....0002等。
这对于查找查询非常有效，现在我添加了排序条件并将_id字段添加到复合索引（见下文）。这在 x100 左右的速度非常糟糕，可能是什么原因？服务器有足够的内存。
db.getCollection('mycoll').find({ p: { '$elemMatch': { k: 'd', v: { '$gt': '2019-10-16T08:01:39.741' } } } }).sort({_id: 1}).explain()
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "mydb.mycoll",
"indexFilterSet" : false,
"parsedQuery" : {
"p" : {
"$elemMatch" : {
"$and" : [
{
"k" : {
"$eq" : "d"
}
},
{
"v" : {
"$gt" : "2019-10-16T08:01:39.741"
}
}
]
}
}
},
"queryHash" : "E28632C1",
"planCacheKey" : "E28632C1",
"winningPlan" : {
"stage" : "SORT",
"sortPattern" : {
"_id" : 1
},
"inputStage" : {
"stage" : "SORT_KEY_GENERATOR",
"inputStage" : {
"stage" : "FETCH",
"filter" : {
"p" : {
"$elemMatch" : {
"$and" : [
{
"k" : {
"$eq" : "d"
}
},
{
"v" : {
"$gt" : "2019-10-16T08:01:39.741"
}
}
]
}
}
},
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"p.k" : 1,
"p.v" : 1,
"_id" : 1
},
"indexName" : "p.k_1_p.v_1__id_1",
"isMultiKey" : true,
"multiKeyPaths" : {
"p.k" : [
"p"
],
"p.v" : [
"p"
],
"_id" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"p.k" : [
"[\"d\", \"d\"]"
],
"p.v" : [
"(\"2019-10-16T08:01:39.741\", {})"
],
"_id" : [
"[MinKey, MaxKey]"
]
}
}
}
}
},
"rejectedPlans" : [
{
"stage" : "FETCH",
"filter" : {
"p" : {
"$elemMatch" : {
"$and" : [
{
"k" : {
"$eq" : "d"
}
},
{
"v" : {
"$gt" : "2019-10-16T08:01:39.741"
}
}
]
}
}
},
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"_id" : 1
},
"indexName" : "_id_",
"isMultiKey" : false,
"multiKeyPaths" : {
"_id" : [ ]
},
"isUnique" : true,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"_id" : [
"[MinKey, MaxKey]"
]
}
}
}
]
},
"serverInfo" : {
"host" : "dev-10-178-3-247",
"port" : 27017,
"version" : "4.2.0",
"gitVersion" : "a4b751dcf51dd249c5865812b390cfd1c0129c30"
},
"ok" : 1
}
> db.getCollection('mycoll').getIndexes()
[
{
"v" : 2,
"key" : {
"_id" : 1
},
"name" : "_id_",
"ns" : "mydb.mycoll"
},
{
"v" : 2,
"key" : {
"p.k" : 1,
"p.v" : 1,
"_id" : 1
},
"name" : "p.k_1_p.v_1__id_1",
"ns" : "mydb.mycoll"
}
]
{
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 1015268,
"executionTimeMillis" : 77032,
"totalKeysExamined" : 3122749,
"totalDocsExamined" : 3122749,
"executionStages" : {
"stage" : "FETCH",
"filter" : {
"p" : {
"$elemMatch" : {
"$and" : [
{
"k" : {
"$eq" : "d"
}
},
{
"v" : {
"$gt" : "2019-10-16T08:01:39.741"
}
}
]
}
}
},
"nReturned" : 1015268,
"executionTimeMillisEstimate" : 16092,
"works" : 3122750,
"advanced" : 1015268,
"needTime" : 2107481,
"needYield" : 0,
"saveState" : 24469,
"restoreState" : 24469,
"isEOF" : 1,
"docsExamined" : 3122749,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 3122749,
"executionTimeMillisEstimate" : 492,
"works" : 3122750,
"advanced" : 3122749,
"needTime" : 0,
"needYield" : 0,
"saveState" : 24469,
"restoreState" : 24469,
"isEOF" : 1,
"keyPattern" : {
"_id" : 1
},
"indexName" : "_id_",
"isMultiKey" : false,
"multiKeyPaths" : {
"_id" : [ ]
},
"isUnique" : true,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"_id" : [
"[MinKey, MaxKey]"
]
},
"keysExamined" : 3122749,
"seeks" : 1,
"dupsTested" : 0,
"dupsDropped" : 0
}
},
"allPlansExecution" : [
{
"nReturned" : 0,
"executionTimeMillisEstimate" : 85,
"totalKeysExamined" : 9298,
"totalDocsExamined" : 9298,
"executionStages" : {
"stage" : "SORT",
"nReturned" : 0,
"executionTimeMillisEstimate" : 85,
"works" : 9300,
"advanced" : 0,
"needTime" : 9299,
"needYield" : 0,
"saveState" : 7391,
"restoreState" : 7391,
"isEOF" : 0,
"sortPattern" : {
"_id" : 1
},
"memUsage" : 33559558,
"memLimit" : 33554432,
"inputStage" : {
"stage" : "SORT_KEY_GENERATOR",
"nReturned" : 9298,
"executionTimeMillisEstimate" : 71,
"works" : 9299,
"advanced" : 9298,
"needTime" : 1,
"needYield" : 0,
"saveState" : 7391,
"restoreState" : 7391,
"isEOF" : 0,
"inputStage" : {
"stage" : "FETCH",
"filter" : {
"p" : {
"$elemMatch" : {
"$and" : [
{
"k" : {
"$eq" : "d"
}
},
{
"v" : {
"$gt" : "2019-10-16T08:01:39.741"
}
}
]
}
}
},
"nReturned" : 9298,
"executionTimeMillisEstimate" : 58,
"works" : 9298,
"advanced" : 9298,
"needTime" : 0,
"needYield" : 0,
"saveState" : 7391,
"restoreState" : 7391,
"isEOF" : 0,
"docsExamined" : 9298,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 9298,
"executionTimeMillisEstimate" : 0,
"works" : 9298,
"advanced" : 9298,
"needTime" : 0,
"needYield" : 0,
"saveState" : 7391,
"restoreState" : 7391,
"isEOF" : 0,
"keyPattern" : {
"p.k" : 1,
"p.v" : 1,
"_id" : 1
},
"indexName" : "p.k_1_p.v_1__id_1",
"isMultiKey" : true,
"multiKeyPaths" : {
"p.k" : [
"p"
],
"p.v" : [
"p"
],
"_id" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"p.k" : [
"[\"d\", \"d\"]"
],
"p.v" : [
"(\"2019-10-16T08:01:39.741\", {})"
],
"_id" : [
"[MinKey, MaxKey]"
]
},
"keysExamined" : 9298,
"seeks" : 1,
"dupsTested" : 9298,
"dupsDropped" : 0
}
}
}
}
},
{
"nReturned" : 0,
"executionTimeMillisEstimate" : 85,
"totalKeysExamined" : 9298,
"totalDocsExamined" : 9298,
"executionStages" : {
"stage" : "SORT",
"nReturned" : 0,
"executionTimeMillisEstimate" : 85,
"works" : 9301,
"advanced" : 0,
"needTime" : 9299,
"needYield" : 0,
"saveState" : 24469,
"restoreState" : 24469,
"isEOF" : 0,
"sortPattern" : {
"_id" : 1
},
"memUsage" : 33559558,
"memLimit" : 33554432,
"inputStage" : {
"stage" : "SORT_KEY_GENERATOR",
"nReturned" : 9298,
"executionTimeMillisEstimate" : 71,
"works" : 9299,
"advanced" : 9298,
"needTime" : 1,
"needYield" : 0,
"saveState" : 24469,
"restoreState" : 24469,
"isEOF" : 0,
"inputStage" : {
"stage" : "FETCH",
"filter" : {
"p" : {
"$elemMatch" : {
"$and" : [
{
"k" : {
"$eq" : "d"
}
},
{
"v" : {
"$gt" : "2019-10-16T08:01:39.741"
}
}
]
}
}
},
"nReturned" : 9298,
"executionTimeMillisEstimate" : 58,
"works" : 9298,
"advanced" : 9298,
"needTime" : 0,
"needYield" : 0,
"saveState" : 24469,
"restoreState" : 24469,
"isEOF" : 0,
"docsExamined" : 9298,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 9298,
"executionTimeMillisEstimate" : 0,
"works" : 9298,
"advanced" : 9298,
"needTime" : 0,
"needYield" : 0,
"saveState" : 24469,
"restoreState" : 24469,
"isEOF" : 0,
"keyPattern" : {
"p.k" : 1,
"p.v" : 1,
"_id" : 1
},
"indexName" : "p.k_1_p.v_1__id_1",
"isMultiKey" : true,
"multiKeyPaths" : {
"p.k" : [
"p"
],
"p.v" : [
"p"
],
"_id" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"p.k" : [
"[\"d\", \"d\"]"
],
"p.v" : [
"(\"2019-10-16T08:01:39.741\", {})"
],
"_id" : [
"[MinKey, MaxKey]"
]
},
"keysExamined" : 9298,
"seeks" : 1,
"dupsTested" : 9298,
"dupsDropped" : 0
}
}
}
}
}
]
},
"ok" : 1
}
解决方案
推荐阅读
- android - 无法在 Android Studio 中重新启动/连接到 ADB：“找不到过程入口点 _create_locale”
- docker - Docker 容器更改默认 DNS 服务器
- spring-boot - 如何使用spring boot在thymleaf中设置聚合对象的输入字段
- angular - 如何从 URL 获取域名 - Angular 7
- scala - Spark 将 TimestampType 转换为格式为 yyyyMMddHHmm 的字符串
- pygame - 树莓派上未显示徽标
- ios - 如何在某个时间线后更改 UITextView 的输入？例如，每 24 小时？迅速
- android - 从导航的深层链接调用我的应用程序会导致我的片段被多次实例化
- canvas - 使用 Fabric.js 中的自定义对象选择功能进行缩放和平移时鼠标偏移的奇怪问题
- django - django manytomanyfield 过滤器