由于文章篇幅有限，均跳過(guò)NoSQL的安裝過(guò)程，請(qǐng)自行參考文檔安裝。

***篇 R利劍MongeDB，分為4個(gè)章節(jié)。

MongoDB環(huán)境準(zhǔn)備
rmongodb函數(shù)庫(kù)
rmongodb基本使用操作
rmongodb測(cè)試案例


每一章節(jié)，都會(huì)分為”文字說(shuō)明部分”和”代碼部分”，保持文字說(shuō)明與代碼的連貫性。

1. MongoDB環(huán)境準(zhǔn)備

文字說(shuō)明部分：

首先環(huán)境準(zhǔn)備，這里我選擇了Linux Ubuntu操作系統(tǒng)12.04的32位桌面版本，大家可以根據(jù)自己的使用習(xí)慣選擇順手的Linux。

MongoDB安裝過(guò)程跳過(guò)。

查看MongoDB服務(wù)器環(huán)境
使用mongod命令，啟動(dòng)MongoDB。

進(jìn)程號(hào)：pid=2924
端口：port=27017
數(shù)據(jù)文件目錄：dbpath=/data/db/
軟件版本：32-bit
主機(jī)名：host=conan


使用mongo命令，打開(kāi)mongo shell。

mongo shell的簡(jiǎn)單操作：
查看數(shù)據(jù)庫(kù)，切換數(shù)據(jù)庫(kù)，查看數(shù)據(jù)集.

R語(yǔ)言環(huán)境2.15.0，WinXP通過(guò)遠(yuǎn)程連接，訪問(wèn)Mongodb Server。

代碼部分：

查看操作系統(tǒng)

~ uname -a

    Linux conan 3.2.0-38-generic-pae #61-Ubuntu SMP Tue Feb 19 12:39:51 UTC 2013 i686 i686 i386 GNU/Linux

~ cat /etc/issue

    Ubuntu 12.04.2 LTS \n \l


啟動(dòng)mongodb

~ mongod

    mongod --help for help and startup options
    Thu Apr 11 11:02:26
    Thu Apr 11 11:02:26 warning: 32-bit servers don't have journaling enabled by default. Please use --journal if you want durability.
    Thu Apr 11 11:02:26
    Thu Apr 11 11:02:26 [initandlisten] MongoDB starting : pid=2924 port=27017 dbpath=/data/db/ 32-bit host=conan
    Thu Apr 11 11:02:26 [initandlisten]
    Thu Apr 11 11:02:26 [initandlisten] ** NOTE: when using MongoDB 32 bit, you are limited to about 2 gigabytes of data
    Thu Apr 11 11:02:26 [initandlisten] **       see http://blog.mongodb.org/post/137788967/32-bit-limitations
    Thu Apr 11 11:02:26 [initandlisten] **       with --journal, the limit is lower
    Thu Apr 11 11:02:26 [initandlisten]
    Thu Apr 11 11:02:26 [initandlisten] db version v2.0.6, pdfile version 4.5
    Thu Apr 11 11:02:26 [initandlisten] git version: e1c0cbc25863f6356aa4e31375add7bb49fb05bc
    Thu Apr 11 11:02:26 [initandlisten] build info: Linux domU-12-31-39-01-70-B4 2.6.21.7-2.fc8xen #1 SMP Fri Feb 15 12:39:36 EST 2008 i686 BOOST_LIB_VERSION=1_41
    Thu Apr 11 11:02:26 [initandlisten] options: {}
    Thu Apr 11 11:02:26 [websvr] admin web console waiting for connections on port 28017
    Thu Apr 11 11:02:26 [initandlisten] waiting for connections on port 27017


打開(kāi)mongo shell

~ mongo

    MongoDB shell version: 2.0.6
    connecting to: test


進(jìn)入mongo shell, 列表顯示數(shù)據(jù)庫(kù)

> show dbs

    db      0.0625GB
    feed    0.0625GB
    foobar  0.0625GB
    local   (empty)


切換數(shù)據(jù)庫(kù)

> use foobar

   switched to db foobar


列表顯示數(shù)據(jù)集

> show collections

    blog
    system.indexes


R語(yǔ)言開(kāi)發(fā)環(huán)境2.15.0，WinXP

~ R
R version 2.15.0 (2012-03-30)
Copyright (C) 2012 The R Foundation for Statistical Computing
ISBN 3-900051-07-0
Platform: i386-pc-mingw32/i386 (32-bit)


2. rmongodb函數(shù)庫(kù)

文字說(shuō)明部分：

rmongodb的開(kāi)發(fā)了一大堆的函數(shù)，對(duì)應(yīng)mongo的操作。比起別的NoSQL來(lái)說(shuō)，真是工程浩大啊。但我總覺(jué)得封裝粒度不夠，寫(xiě)起代碼來(lái)比較復(fù)雜。

下面列出了所有rmongodb函數(shù)庫(kù)，我只挑選幾個(gè)常用的介紹。

建立mongo連接

mongo<-mongo.create()


查看接連是否正常

mongo.is.connected(mongo)


創(chuàng)建一個(gè)BSON對(duì)象緩存

buf <- mongo.bson.buffer.create()


給對(duì)象buf增加element

mongo.bson.buffer.append(buf, "name", "Echo")


增加對(duì)象類(lèi)型的element

score <- c(5, 3.5, 4)
names(score) <- c("Mike", "Jimmy", "Ann")
mongo.bson.buffer.append(buf, "score", score)


增加數(shù)組類(lèi)型的element

mongo.bson.buffer.start.array(buf, "comments")
mongo.bson.buffer.append(buf, "0", "a1")
mongo.bson.buffer.append(buf, "1", "a2")
mongo.bson.buffer.append(buf, "2", "a3")


關(guān)閉數(shù)組類(lèi)型的element

mongo.bson.buffer.finish.object(buf)


取出緩存數(shù)據(jù)

b <- mongo.bson.from.buffer(buf)


數(shù)據(jù)庫(kù).數(shù)據(jù)集

ns="db.blog"


插入一條記錄

mongo.insert(mongo,ns,b)

#mongo shell:(Not Run)
db.blog.insert(b)


創(chuàng)建查詢(xún)對(duì)象query

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.append(buf, "name", "Echo")
query <- mongo.bson.from.buffer(buf)


創(chuàng)建查詢(xún)返回值對(duì)象

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.append(buf, "name", 1)
fields <- mongo.bson.from.buffer(buf)


執(zhí)行單條記錄查詢(xún)

mongo.find.one(mongo, ns, query, fields)

#mongo shell:(Not Run)
db.blog.findOne({query},{fields})


執(zhí)行列表記錄查詢(xún)

mongo.find(mongo, ns, query, fields)

#mongo shell:(Not Run)
db.blog.find({query},{fields})


創(chuàng)建修改器對(duì)象objNew

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.start.object(buf, "$inc")
mongo.bson.buffer.append(buf, "age", 1L)
mongo.bson.buffer.finish.object(buf)
objNew <- mongo.bson.from.buffer(buf)


執(zhí)行修改操作

mongo.update(mongo, ns, query, objNew)

#mongo shell:(Not Run)
db.blog.update({query},{objNew})


單行代碼修改操作

mongo.update(mongo, ns, query, list(name="Echo", age=25))

#mongo shell:(Not Run)
db.blog.update({query},{objNew})


刪除所選對(duì)象

mongo.remove(mongo, ns, query)

#mongo shell:(Not Run)
db.blog.remove({query},{objNew})


銷(xiāo)毀mongo連接

mongo.destroy(mongo)


代碼部分：

共有153個(gè)函數(shù)

mongo.add.user
mongo.authenticate
mongo.binary.binary
mongo.binary.function
mongo.binary.md5
mongo.binary.old
mongo.binary.user
mongo.binary.uuid
mongo.bson.array
mongo.bson.binary
mongo.bson.bool
mongo.bson.buffer.append
mongo.bson.buffer.append.bool
mongo.bson.buffer.append.bson
mongo.bson.buffer.append.code
mongo.bson.buffer.append.code.w.scope
mongo.bson.buffer.append.complex
mongo.bson.buffer.append.double
mongo.bson.buffer.append.element
mongo.bson.buffer.append.int
mongo.bson.buffer.append.list
mongo.bson.buffer.append.long
mongo.bson.buffer.append.null
mongo.bson.buffer.append.object
mongo.bson.buffer.append.oid
mongo.bson.buffer.append.raw
mongo.bson.buffer.append.regex
mongo.bson.buffer.append.string
mongo.bson.buffer.append.symbol
mongo.bson.buffer.append.time
mongo.bson.buffer.append.timestamp
mongo.bson.buffer.append.undefined
mongo.bson.buffer.create
mongo.bson.buffer.finish.object
mongo.bson.buffer.size
mongo.bson.buffer.start.array
mongo.bson.buffer.start.object
mongo.bson.code
mongo.bson.code.w.scope
mongo.bson.date
mongo.bson.dbref
mongo.bson.destroy
mongo.bson.double
mongo.bson.empty
mongo.bson.eoo
mongo.bson.find
mongo.bson.from.buffer
mongo.bson.from.list
mongo.bson.int
mongo.bson.iterator.create
mongo.bson.iterator.key
mongo.bson.iterator.next
mongo.bson.iterator.type
mongo.bson.iterator.value
mongo.bson.long
mongo.bson.null
mongo.bson.object
mongo.bson.oid
mongo.bson.print
mongo.bson.regex
mongo.bson.size
mongo.bson.string
mongo.bson.symbol
mongo.bson.timestamp
mongo.bson.to.list
mongo.bson.undefined
mongo.bson.value
mongo.code.create
mongo.code.w.scope.create
mongo.command
mongo.count
mongo.create
mongo.cursor.destroy
mongo.cursor.next
mongo.cursor.value
mongo.destroy
mongo.disconnect
mongo.distinct
mongo.drop
mongo.drop.database
mongo.find
mongo.find.await.data
mongo.find.cursor.tailable
mongo.find.exhaust
mongo.find.no.cursor.timeout
mongo.find.one
mongo.find.oplog.replay
mongo.find.partial.results
mongo.find.slave.ok
mongo.get.database.collections
mongo.get.databases
mongo.get.err
mongo.get.hosts
mongo.get.last.err
mongo.get.prev.err
mongo.get.primary
mongo.get.server.err
mongo.get.server.err.string
mongo.get.socket
mongo.get.timeout
mongo.gridfile.destroy
mongo.gridfile.get.chunk
mongo.gridfile.get.chunk.count
mongo.gridfile.get.chunks
mongo.gridfile.get.chunk.size
mongo.gridfile.get.content.type
mongo.gridfile.get.descriptor
mongo.gridfile.get.filename
mongo.gridfile.get.length
mongo.gridfile.get.md5
mongo.gridfile.get.metadata
mongo.gridfile.get.upload.date
mongo.gridfile.pipe
mongo.gridfile.read
mongo.gridfile.seek
mongo.gridfile.writer.create
mongo.gridfile.writer.finish
mongo.gridfile.writer.write
mongo.gridfs.create
mongo.gridfs.destroy
mongo.gridfs.find
mongo.gridfs.remove.file
mongo.gridfs.store
mongo.gridfs.store.file
mongo.index.background
mongo.index.create
mongo.index.drop.dups
mongo.index.sparse
mongo.index.unique
mongo.insert
mongo.insert.batch
mongo.is.connected
mongo.is.master
mongo.oid.create
mongo.oid.from.string
mongo.oid.print
mongo.oid.time
mongo.oid.to.string
mongo.reconnect
mongo.regex.create
mongo.remove
mongo.rename
mongo.reset.err
mongo.set.timeout
mongo.shorthand
mongo.simple.command
mongo.symbol.create
mongo.timestamp.create
mongo.undefined.create
mongo.update
mongo.update.basic
mongo.update.multi
mongo.update.upsert


3. rmongodb基本使用操作

文字說(shuō)明部分：

首先，要安裝rmongodb類(lèi)庫(kù)，加載類(lèi)庫(kù)。

然后，通過(guò)mongo.create()函數(shù)，建立與MongoDB Server的連接。如果是本地連接，mongo.create()不要參數(shù)，下面例子使用遠(yuǎn)程連接，增加host參數(shù)配置IP地址。mongo<-mongo.create(host=“192.168.1.11”)

檢查是否連接正常，mongo.is.connected()。這條語(yǔ)句在開(kāi)發(fā)時(shí)會(huì)經(jīng)常使用到。在用R語(yǔ)言建模時(shí)，如果對(duì)象或者函數(shù)使用錯(cuò)誤，連接會(huì)被自動(dòng)斷開(kāi)。由于MongoDB的異常機(jī)制，斷開(kāi)時(shí)不會(huì)是提示。大家要手動(dòng)使用這條命令測(cè)試，連接是否正常。

接下來(lái)，定義兩個(gè)變量，db和ns。db是我們需要使用的數(shù)據(jù)庫(kù)，ns是數(shù)據(jù)庫(kù)+數(shù)據(jù)集。

下面我們創(chuàng)建一個(gè)Mongo對(duì)象。

{
        "_id" : ObjectId("51663e14da2c51b1e8bc62eb"),
        "name" : "Echo",
        "age" : 22,
        "gender" : "Male",
        "score" : {
                "Mike" : 5,
                "Jimmy" : 3.5,
                "Ann" : 4
        },
        "comments" : [
                "a1",
                "a2",
                "a3"
        ]
}


然后，分別使用修改器inc,set,$push進(jìn)行操作。

***刪除對(duì)象，并斷開(kāi)連接。

代碼部分：

安裝rmongodb

install.packages(rmongodb)


加載類(lèi)庫(kù)

library(rmongodb)


遠(yuǎn)程連接mongodb server

mongo<-mongo.create(host="192.168.1.11")


查看是否連接正常

print(mongo.is.connected(mongo))


定義db

db<-"foobar"


定義db.collection

ns<-"foobar.blog"


組織bson類(lèi)型數(shù)據(jù)

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.append(buf, "name", "Echo")
mongo.bson.buffer.append(buf, "age", 22L)
mongo.bson.buffer.append(buf, "gender", 'Male')

#對(duì)象類(lèi)型
score <- c(5, 3.5, 4)
names(score) <- c("Mike", "Jimmy", "Ann")
mongo.bson.buffer.append(buf, "score", score)

#數(shù)組類(lèi)型
mongo.bson.buffer.start.array(buf, "comments")
mongo.bson.buffer.append(buf, "0", "a1")
mongo.bson.buffer.append(buf, "1", "a2")
mongo.bson.buffer.append(buf, "2", "a3")
mongo.bson.buffer.finish.object(buf)
b <- mongo.bson.from.buffer(buf)


插入mongodb

mongo.insert(mongo,ns,b)


單條顯示插入的數(shù)據(jù)

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.append(buf, "name", "Echo")
query <- mongo.bson.from.buffer(buf)
print(mongo.find.one(mongo, ns, query))


使用$inc修改器，修改給age加1

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.start.object(buf, "$inc")
mongo.bson.buffer.append(buf, "age", 1L)
mongo.bson.buffer.finish.object(buf)
objNew <- mongo.bson.from.buffer(buf)
mongo.update(mongo, ns, query, objNew)
print(mongo.find.one(mongo, ns, query))


使用$set修改器，修改age=1

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.start.object(buf, "$set")
mongo.bson.buffer.append(buf, "age", 1L)
mongo.bson.buffer.finish.object(buf)
objNew <- mongo.bson.from.buffer(buf)
mongo.update(mongo, ns, query, objNew)
print(mongo.find.one(mongo, ns, query))


使用$push修改器，給comments數(shù)組追加”Orange”數(shù)據(jù)

buf <- mongo.bson.buffer.create()
mongo.bson.buffer.start.object(buf, "$push")
mongo.bson.buffer.append(buf, "comments", "Orange")
mongo.bson.buffer.finish.object(buf)
objNew <- mongo.bson.from.buffer(buf)
mongo.update(mongo, ns, query, objNew)
print(mongo.find.one(mongo, ns, query))


使用簡(jiǎn)化修改語(yǔ)句，給對(duì)象重新賦值

mongo.update(mongo, ns, query, list(name="Echo", age=25))
print(mongo.find.one(mongo, ns, query))


刪除對(duì)象

mongo.remove(mongo, ns, query)


銷(xiāo)毀mongo連接

mongo.destroy(mongo)


4. rmongodb測(cè)試案例

文字說(shuō)明部分：

批量插入數(shù)據(jù)，使用修改器批量修改數(shù)據(jù)

3種修改器速度比較,push最慢push > set>inc

終于push是對(duì)數(shù)組操作，set是對(duì)任意值操作，inc是對(duì)數(shù)字操作，所以下面測(cè)試可能不太公平。測(cè)試結(jié)果僅供參考。

代碼部分：

批量插入數(shù)據(jù)函數(shù)

  batch_insert<-function(arr=1:10,ns){
    library(stringr)
    mongo_insert<-function(x){
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.append(buf, "name", str_c("Dave",x))
      mongo.bson.buffer.append(buf, "age", x)
      mongo.bson.buffer.start.array(buf, "comments")
      mongo.bson.buffer.append(buf, "0", "a1")
      mongo.bson.buffer.append(buf, "1", "a2")
      mongo.bson.buffer.append(buf, "2", "a3")
      mongo.bson.buffer.finish.object(buf)
      return(mongo.bson.from.buffer(buf))
    }
    mongo.insert.batch(mongo, ns, lapply(arr,mongo_insert))
  }


批量修改，$inc修改器函數(shù)

  batch_inc<-function(data,ns){
    for(i in data){
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.append(buf, "name", str_c("Dave",i))
      criteria <- mongo.bson.from.buffer(buf)
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.start.object(buf, "$inc")
      mongo.bson.buffer.append(buf, "age", 1L)
      mongo.bson.buffer.finish.object(buf)
      objNew <- mongo.bson.from.buffer(buf)
      mongo.update(mongo, ns, criteria, objNew)
    }
  }


批量修改，$set修改器函數(shù)

  batch_set<-function(data,ns){
    for(i in data){
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.append(buf, "name", str_c("Dave",i))
      criteria <- mongo.bson.from.buffer(buf)
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.start.object(buf, "$set")
      mongo.bson.buffer.append(buf, "age", 1L)
      mongo.bson.buffer.finish.object(buf)
      objNew <- mongo.bson.from.buffer(buf)
      mongo.update(mongo, ns, criteria, objNew)
    }
  }


批量修改，$push修改器函數(shù)

  batch_push<-function(data,ns){
    for(i in data){
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.append(buf, "name", str_c("Dave",i))
      criteria <- mongo.bson.from.buffer(buf)
      buf <- mongo.bson.buffer.create()
      mongo.bson.buffer.start.object(buf, "$push")
      mongo.bson.buffer.append(buf, "comments", "Orange")
      mongo.bson.buffer.finish.object(buf)
      objNew <- mongo.bson.from.buffer(buf)
      mongo.update(mongo, ns, criteria, objNew)
    }
  }


執(zhí)行程序，3種修改速度比較,$push最慢

ns="foobar.blog" data=1:1000 mongo.remove(mongo, ns) ## [1] TRUE system.time(batch_insert(data, ns)) ## user system elapsed ## 0.25 0.00 0.28 system.time(batch_inc(data, ns)) ## user system elapsed ## 0.47 0.27 2.50 system.time(batch_set(data, ns)) ## user system elapsed ## 0.77 0.48 3.17 system.time(batch_push(data, ns)) ## user system elapsed ## 0.81 0.41 4.23