MongoDB-基本功能

目标：



掌握mongodb的基本功能

• 用数据填充MongoDB;
• 编写多个查询以获取和更新数据
• 创建索引并比较性能。

填充数据仓库



为了寻找足够的数据来研究MongoDB的基本功能，我选择了NASA地球陨石着陆数据集（从存储库https://github.com/jdorfman/awesome-收集了1000行有关落到地球的陨石的信息）JSON-数据集。



注：发现较完整的（45.7K）信息https://data.nasa.gov/Space-Science/Meteorite-Landings/gh4g-9sfh，但通过他们的API出口到JSON只给出了1000条记录（不知所云） ，就必须铲导出CSV文件的完整的数据https://data.nasa.gov/api/views/gh4g-9sfh/rows.csv?accessType=DOWNLOAD？



注：糟糕，可以使用JSON获取完整的数据，但这是一个技巧。我衷心希望这不是SQL注入



https :? //Data.nasa.gov/api/id/gh4g-9sfh.json $ select =` name`，`id`，`nametype`，`recclass`，`质量`，'下降`，`年`，`reclat`，`reclong`，`geolocation`和$ order =`：id` + ASC＆$ limit = 46000＆$ offset = 0

wc ./gh4g-9sfh.json
     45716   128491 10441343 ./gh4g-9sfh.json

为了与网络上的服务器进行交互，我仅在本地计算机上安装了客户端和工具：

wget -qO - https://www.mongodb.org/static/pgp/server-4.4.asc | sudo apt-key add -
echo "deb http://repo.mongodb.org/apt/debian buster/mongodb-org/4.4 main" | sudo tee /etc/apt/sources.list.d/mongodb-org-4.4.list

sudo apt-get update
sudo apt-get install -y mongodb-org-shell
sudo apt-get install -y mongodb-org-tools

连接检查：

mongo nosql-2020.otus --port 32789

    MongoDB shell version v4.4.1
    connecting to: mongodb://nosql-2020.otus:32789/test?compressors=disabled&gssapiServiceName=mongodb
    Implicit session: session { "id" : UUID("5ff24788-0710-4a1a-821f-7acb2eddfb4f") }
    MongoDB server version: 4.4.1
    Welcome to the MongoDB shell.
    For interactive help, type "help".
    For more comprehensive documentation, see
            https://docs.mongodb.com/
    Questions? Try the MongoDB Developer Community Forums
            https://community.mongodb.com

将数据从本地计算机导入到远程计算机：

mongoimport --host nosql-2020.otus  --db "otus_003" --port 32789 --collection eml45k --jsonArray --file ./003_MONGODB.files/gh4g-9sfh.json
    
    2020-10-12T01:01:13.826+0100    connected to: mongodb://nosql-2020.otus:32789/
    2020-10-12T01:01:16.827+0100    [#######.................] otus_003.eml45k      2.99MB/9.96MB (30.0%)
    2020-10-12T01:01:19.827+0100    [###############.........] otus_003.eml45k      6.44MB/9.96MB (64.6%)
    2020-10-12T01:01:22.827+0100    [#######################.] otus_003.eml45k      9.81MB/9.96MB (98.5%)
    2020-10-12T01:01:23.035+0100    [########################] otus_003.eml45k      9.96MB/9.96MB (100.0%)
    2020-10-12T01:01:23.035+0100    45716 document(s) imported successfully. 0 document(s) failed to import.

注意：所有内容需要10秒，有趣的数据提取

> show databases
admin     0.000GB
config    0.000GB
local     0.000GB
otus_003  0.000GB
test      0.000GB
> use otus_003
switched to db otus_003
> show collections
em
l

我们正在寻找一个名称已知的陨石：

> db.eml45k.find({name:"Bjelaja Zerkov"})

    { "_id" : ObjectId("5f8380a91c0ab84b54bfe394"), "name" : "Bjelaja Zerkov", "id" : "5063", "nametype" : "Valid", "recclass" : "H6", "mass" : "1850", "fall" : "Fell", "year" : "1796-01-01T00:00:00.000", "reclat" : "49.783330", "reclong" : "30.166670", "geolocation" : { "latitude" : "49.78333", "longitude" : "30.16667" } }

我们正在寻找使用已知坐标的陨石：

> db.eml45k.find({ "geolocation" : { "latitude" : "44.83333" , "longitude" : "95.16667" } })

{ "_id" : ObjectId("5f8380a91c0ab84b54bfe322"), "name" : "Adzhi-Bogdo (stone)", "id" : "390", "nametype" : "Valid", "recclass" : "LL3-6", "mass" : "910", "fall" : "Fell", "year" : "1949-01-01T00:00:00.000", "reclat" : "44.833330", "reclong" : "95.166670", "geolocation" : { "latitude" : "44.83333", "longitude" : "95.16667" } }

获取按坠落年份排序的坠落陨石列表（我想知道为什么美国航空航天局在格林威治标准时间没有特定的坠落时间），并且具有有限的可选字段列表：

> db.eml45k.find( { }, {year: 1, id: 1, name: 1, _id: 0 }).sort( { year: -1 } )

{ "name" : "Northwest Africa 7701", "id" : "57150", "year" : "2101-01-01T00:00:00.000" }
{ "name" : "Chelyabinsk", "id" : "57165", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7755", "id" : "57166", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7812", "id" : "57258", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7822", "id" : "57268", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7856", "id" : "57421", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7855", "id" : "57420", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7857", "id" : "57422", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7858", "id" : "57423", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7861", "id" : "57425", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7862", "id" : "57426", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Northwest Africa 7863", "id" : "57427", "year" : "2013-01-01T00:00:00.000" }
{ "name" : "Battle Mountain", "id" : "56133", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Sutter's Mill", "id" : "55529", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Antelope", "id" : "57455", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Catalina 009", "id" : "57173", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Jiddat al Harasis 799", "id" : "57428", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Johannesburg", "id" : "55765", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Ksar Ghilane 011", "id" : "55606", "year" : "2012-01-01T00:00:00.000" }
{ "name" : "Ksar Ghilane 010", "id" : "55605", "year" : "2012-01-01T00:00:00.000" }
Type "it" for more
>

我还没找到？如何在选择时直接合并两个字段：

"geolocation" : { "latitude" : "49.78333", "longitude" : "30.16667" } 

到一定程度（应注意，顺序已根据文档进行了更改（https://docs.mongodb.com/manual/geospatial-queries/#geospatial-legacy，即

< field >: [< longitude >, < latitude >]):

"geolocation" : { "type" : "Point", "coordinates" : [  30.16667 , 49.78333 ] } }

直接被要求做（谁知道？）这样的事情：

> db.eml45k.find({ 
    [ 
        {$toDouble: "$geolocation.longitude"} ,
        {$toDouble: "$geolocation.latitude"} 
    ] : {
        $geoWithin: {
            $geometry: {
                type : "Polygon" ,
                coordinates: [[
                    ... ,
                ]]
            }
        }
    }
}) 

因此，我在集合中创建了一个人工字段：

db.eml45k.updateMany( 
    {},
    [{
        $set: {
            "pointed_geolocation.type" : "Point",
            "pointed_geolocation.coordinates" : [ 
                { $toDouble : "$geolocation.longitude" } , 
                { $toDouble: "$geolocation.latitude" } 
            ]
        }
    }]
);

{ "acknowledged" : true, "matchedCount" : 45716, "modifiedCount" : 45716 }

最后我们可以寻找落在某个区域的未检测到的陨石：

> db.eml45k.find({ 
    "pointed_geolocation.coordinates" : {
        $geoWithin: {
            $geometry: {
                type : "Polygon" ,
                coordinates: [[
                    [ 47.0 , 33.0  ], 
                    [ 47.0 , 65.0 ], 
                    [ 169.0 , 65.0 ],
                    [ 169.0 ,  33.0 ],
                    [ 47.0 , 33.0 ]
                ]]
            }
        }
    },
    'fall': 'Fell'
},
{ 
    year: {$year: { "$toDate": "$year"}}, 
    "pointed_geolocation.coordinates": 1, 
    name: 1, 
    _id: 0 
}).sort( { year: -1 } )

奇怪，我不知道车里雅宾斯基没有在找到的类别中列出。



让我们合计并发现有多少未找到：

db.eml45k.aggregate([
{ $match: { 
    "pointed_geolocation.coordinates" : {
        $geoWithin: {
            $geometry: {
                type : "Polygon" ,
                coordinates: [[
                    [ 47.0 , 33.0  ], 
                    [ 47.0 , 65.0 ], 
                    [ 169.0 , 65.0 ],
                    [ 169.0 ,  33.0 ],
                    [ 47.0 , 33.0 ]
                ]]
            }
        }
    }
} },
    {"$group" : {_id: "$fall", count: { $sum: 1 }}}
])

{ "_id" : "Fell", "count" : 26 }
{ "_id" : "Found", "count" : 63 }

总共发现了89个中的63个，并且找到了__26__-__not__，所以有机会:)



使用索引



让我们从以前的实验中删除集合中的所有索引：

db.eml45k.dropIndexes()

{
        "nIndexesWas" : 1,
        "msg" : "non-_id indexes dropped for collection",
        "ok" : 1
}

让我们尝试查看请求的预计执行时间：

db.eml45k.find({ 
    "pointed_geolocation.coordinates" : {
        $geoWithin: {
            $geometry: {
                type : "Polygon" ,
                coordinates: [[
                    [ 47.0 , 33.0  ], 
                    [ 47.0 , 65.0 ], 
                    [ 169.0 , 65.0 ],
                    [ 169.0 ,  33.0 ],
                    [ 47.0 , 33.0 ]
                ]]
            }
        }
    }
}).explain("executionStats").executionStats.executionTimeMillis
...
110
...
110
...
109

结果平均约为110秒。



索引：

db.eml45k.createIndex( { "pointed_geolocation" : "2dsphere" } )

    {
        "ok" : 0,
        "errmsg" : "Index build failed: 98b9ead2-c156-4312-81af-1adf5896e3c9: Collection otus_003.eml45k ( 6db2d178-61b5-4627-8512-fcf919fe596f ) :: caused by :: Can't extract geo keys: { _id: ObjectId('5f838e30fb89bd9d553ae27f'), name: \"Bulls Run\", id: \"5163\", nametype: \"Valid\", recclass: \"Iron?\", mass: \"2250\", fall: \"Fell\", year: \"1964-01-01T00:00:00.000\", pointed_geolocation: { type: \"Point\", coordinates: [ null, null ] } }  Point must only contain numeric elements",
        "code" : 16755,
        "codeName" : "Location16755"
    }

该错误是由于NULL值引起的，我没有立即找到（知道吗？）如何在索引期间将其从索引中排除，所以我将删除这些键：

db.eml45k.updateMany(
    { "pointed_geolocation.coordinates" : [ null , null ] },     
    [{         
        $set: { "pointed_geolocation": null }
    }] 
);

再次尝试索引

db.eml45k.createIndex( { "pointed_geolocation" : "2dsphere" } )
    
    {
        "ok" : 0,
        "errmsg" : "Index build failed: d33b31d4-4778-4537-a087-58b7bd1968f3: Collection otus_003.eml45k ( 6db2d178-61b5-4627-8512-fcf919fe596f ) :: caused by :: Can't extract geo keys: { _id: ObjectId('5f838e35fb89bd9d553b3b8f'), name: \"Meridiani Planum\", id: \"32789\", nametype: \"Valid\", recclass: \"Iron, IAB complex\", fall: \"Found\", year: \"2005-01-01T00:00:00.000\", reclat: \"-1.946170\", reclong: \"354.473330\", geolocation: { latitude: \"-1.94617\", longitude: \"354.47333\" }, pointed_geolocation: { type: \"Point\", coordinates: [ 354.47333, -1.94617 ] } }  longitude/latitude is out of bounds, lng: 354.473 lat: -1.94617",
        "code" : 16755,
        "codeName" : "Location16755"
    }

错误__经度/纬度超出范围，lng：354.473 lat：-1.94617__和文档https://docs.mongodb.com/manual/geospatial-queries/#geospatial-legacy

    Valid longitude values are between -180 and 180, both inclusive.
    Valid latitude values are between -90 and 90, both inclusive.

而354.47333不在-180到180的范围内。



非常奇怪，起初我认为到处都需要在负180处进行修正以使

(`$subtract: [{ $toDouble : "$geolocation.longitude" }, 180.0]`)

，但最终一切都不是那么简单。



哪些经度不在范围内：

db.eml45k.find({"pointed_geolocation.coordinates.0": {$lt: -180}} ) #  ,    
db.eml45k.find({"pointed_geolocation.coordinates.0": {$lt: 0}} ) #   ,    
db.eml45k.find({"pointed_geolocation.coordinates.0": {$gt: 180}} ) #  ,     


    { "_id" : ObjectId("5f838e35fb89bd9d553b3b8f"), "name" : "Meridiani Planum", "id" : "32789", "nametype" : "Valid", "recclass" : "Iron, IAB complex", "fall" : "Found", "year" : "2005-01-01T00:00:00.000", "reclat" : "-1.946170", "reclong" : "354.473330", "geolocation" : { "latitude" : "-1.94617", "longitude" : "354.47333" }, "pointed_geolocation" : { "type" : "Point", "coordinates" : [ 354.47333, -1.94617 ] } }

结果，只有一个陨石具有奇怪的坐标。搜索后，我发现此Meridiani Planum陨石是2005年由商机意外发现的

（http://old.mirf.ru/Articles/art2427_2.htm）。这是一个（警告）火星陨石;实测值（警告）在火星上。这是美国宇航局的笑话。



让我们从集合中删除它。

db.eml45k.remove({"id" : "32789"})
WriteResult({ "nRemoved" : 1 })

索引编制

> db.eml45k.createIndex( { "pointed_geolocation" : "2dsphere" } )
{
        "createdCollectionAutomatically" : false,
        "numIndexesBefore" : 1,
        "numIndexesAfter" : 2,
        "ok" : 1
}

我们测量

db.eml45k.find({ 
    "pointed_geolocation.coordinates" : {
        $geoWithin: {
            $geometry: {
                type : "Polygon" ,
                coordinates: [[
                    [ 47.0 , 33.0  ], 
                    [ 47.0 , 65.0 ], 
                    [ 169.0 , 65.0 ],
                    [ 169.0 ,  33.0 ],
                    [ 47.0 , 33.0 ]
                ]]
            }
        }
    }
}).explain("executionStats").executionStats.executionTimeMillis

结果，当重新检查104 ... 107 ... 106 ...时，



这有点奇怪，不是很明亮。



删除索引，选中。



没有索引但有索引-相同。



我为车里雅宾斯克单独尝试：

db.eml45k.find(
    {"pointed_geolocation.coordinates" : [ 61.11667, 54.81667 ]}
).explain("executionStats").executionStats.executionTimeMillis

没有索引，但有索引-相同。



你需要更加小心，该指数是内置pointed_geolocation场，并pointed_geolocation.coordinates在查询中涉及的。



其结果是，查询

db.eml45k.find({ 
    "pointed_geolocation" : {
        $geoWithin: {
            $geometry: {
                type : "Polygon" ,
                coordinates: [[
                    [ 47.0 , 33.0  ], 
                    [ 47.0 , 65.0 ], 
                    [ 169.0 , 65.0 ],
                    [ 169.0 ,  33.0 ],
                    [ 47.0 , 33.0 ]
                ]]
            }
        }
    }
}).explain("executionStats").executionStats.executionTimeMillis

没有索引125、123、119、123毫秒，而索引为



7、4、4、5。