从批处理文件执行 rga.open(instance="ga", where="~/ga.rga") 时出错答案

【问题标题】：Error on executing rga.open(instance="ga", where="~/ga.rga") from a batch file从批处理文件执行 rga.open(instance="ga", where="~/ga.rga") 时出错
【发布时间】：2014-08-07 16:30:45
【问题描述】：

我正在尝试通过批处理文件运行 r 脚本，以便我可以安排 R 脚本定期运行。 R 脚本从 Google Analytics 检索数据并将数据转储到 SQL Server 数据库中。

R 脚本在 R Studio 下运行得非常好，但是当我尝试通过批处理文件执行相同的 R 脚本时，它会出现此错误 -

“body_config(body, multipart) 中的错误：找不到对象 redirect.uri 调用 rga.open ... modify_config -> setdiff -> as.vector -> body_config 执行停止”

我的批处理文件如下所示： "C:\Program Files\R\R-3.0.2\bin\Rscript.exe" "C:\Users\sc\Documents\R Sources\SocialMedia\ExtractBlogStats.r" 暂停

我的 Rscript 看起来像：

library(devtools)

#install_github("rga", "skardhamar")
library(rga)

#install.packages("RODBC")
library(RODBC)

myconn <- odbcDriverConnect('driver={SQL Server};server=localhost;database=SocialMedia;trusted_connection=true')


options(RCurlOptions = list(cainfo = system.file("CurlSSL", "cacert.pem", package = "RCurl")))

#rga.open(instance="ga")


rga.open(instance="ga", where="~/ga.rga")


id <- "XXXXXXXX" # Valid ID in actual script file

end <- Sys.Date() - 1
start <- end - 1

by.practice <- ga$getData(id, start.date = start, end.date= end, 
                        metrics = "ga:sessions,ga:totalEvents",
                        dimensions = "ga:eventAction",
                        sort = "",
                        filters = "",
                        segment = "",
                        start = 1,
                        max = 1000)

【问题讨论】：

这里只是一个疯狂的猜测，但你是在代理后面工作吗？这可以解释为什么它可以在 RStudio 中工作而不是作为批处理脚本调用。
您是否在运行 RStudio 的同一台机器上运行批处理文件？并且作为同一个用户？路径中的“~”用于用户特定的主目录，如果批处理脚本由调度程序运行，则不同的用户主目录可能用于“~”。也许尝试在那里设置完整路径？
@MrFlick - 是的，我在运行 RStudio 的同一台机器上以同一用户身份运行批处理文件。我的批处理文件中的两个路径都是完整路径。你有什么建议吗？
@user3749083 看起来where="~/ga.rga" 可能会因为那个波浪号而变得麻烦，并且在一个地方它使用了缓存的身份验证，而在另一个地方却没有。只是为了检查您可以打印出path.expand("~/ga.rga") 以查看它的解析位置。如果您在rga.open 之后仍然收到错误，您可能还需要添加traceback() 以了解它当时正在尝试做什么。
@MrFlick - 我包含了 path.expand 并注意到当我在 RStudio ("C:/Users/sc/Documents/ga.rga") 与批处理文件中运行时值存在差异（“C:\\Users\\sc\\Documents/ga.rga”）。我阅读了 where 子句并发现“如果您想在本地存储实例，可以通过添加 where 属性来完成：”。所以我修改了我的代码以指定显式路径而不是波浪号 [where ="C:/Users/sc/Documents/R Sources/SocialMedia/ga.rga"] 但它没有帮助。另外，我添加了 setwd("C:/Users/sc/Documents/R Sources/SocialMedia") 但这也无济于事。

标签： r batch-file google-analytics

【解决方案1】：

步骤： a) 在 RStudio 中执行 rga.open(instance="ga")

b) 这将我引导至 URL 并生成代码

c) 将代码粘贴到 RStudio 控制台中

d) 本地文件系统上保存的 ga 对象 -- save(ga, file="C:/Users/sc/Documents/R 来源/SocialMedia/ga.rga")

e) 从函数中删除了 rga.open 调用 - extract.blog.stats.by.practice()

extract.blog.stats.by.practice = function(id = "XXXXXXX") {
library(devtools)
library(rga)
library(RODBC)
myconn <- odbcDriverConnect('driver={SQL Server};server=localhost;database=SocialMediaSink;trusted_connection=true')

options(RCurlOptions = list(cainfo = system.file("CurlSSL", "cacert.pem", package = "RCurl")))

#rga.open(instance="ga")
setwd("C:/Users/sc/Documents/R Sources/SocialMedia")

end <- Sys.Date() - 1
start <- end - 1

by.practice <- NULL

by.practice <- ga$getData(id, start.date = start, end.date= end, 
                        metrics = "ga:sessions,ga:totalEvents",
                        dimensions = "ga:eventAction",
                        sort = "",
                        filters = "",
                        segment = "",
                        start = 1,
                        max = 1000)


# Convert Numeric To Integer
practice <- by.practice[,1]
sessions <- as.integer(by.practice[,2])
totalEvents <- as.integer(by.practice[,3])

# Create a dataframe that mimics the structure of the database table - BlogPracticeStats
practice.stats <- data.frame(practice, end, sessions, totalEvents, stringsAsFactors = FALSE)

# Assign columnnames that match the table's column name in the database
colnames(practice.stats) <- c("Practice", "Date", "Sessions", "TotalEvents")

# Insert rows in the table
sqlSave(myconn, practice.stats, "BlogPracticeStats", safer = FALSE, append = TRUE, rownames = FALSE)
odbcClose(myconn)

 }

f) 在 RunFunction.r 中包含来自本地保存副本的“ga”负载

source("C:/Users/sc/Documents/R Sources/SocialMedia/ExtractBlogPracticeStats.R")
load("C:/Users/sc/Documents/R Sources/SocialMedia/ga.rga")
extract.blog.stats.by.practice()

g) 执行的批处理文件

"C:\Program Files\R\R-3.0.2\bin\Rscript.exe" "C:\Users\sc\Documents\R Sources\SocialMedia\RunFunction.r"
pause

【讨论】：

【解决方案2】：

我想发布整个解决方案 -

第 1 步——创建初始身份验证对象

#Initial Objects
# dev tools is needed because we need to use it's function install_github

# install.packages("devtools")
library(devtools)

# install_github("rga", "skardhamar")
library(rga)

library(RODBC)

config.folder.location <- "C:\\Users\\sc\\Documents\\R Sources\\SocialMedia\\Config"

working.directory <- config.folder.location

rga.open(instance="ga")

save(ga, file="C:/Users/sc/Documents/R Sources/SocialMedia/config/ga.rga")


connection.string <- 'driver={SQL Server};server=localhost;database=SocialMedia;trusted_connection=true'

save(connection.string,  file = paste0(config.folder.location,'\\connection.string '))


# Google Analytics ID for ur site
google.analytics.id <- "XXXXXX"

save(google.analytics.id, file="google.analytics.id")

第 2 步 -- 查询 Google Analaytics

# dev tools is needed because we need to use it's function install_github
# install.packages("devtools")
library(devtools)

# install_github("rga", "skardhamar")
library(rga)

# install.packages("RODBC")
library(RODBC)


#*********************************************************************************************************************
# Read arguments and store them in variables -- START
#*********************************************************************************************************************
args <- commandArgs(trailingOnly = TRUE)

# First argument is read as working directory (location where config folder resides)
working.directory <- as.character(args[1])

if (length(args) >= 2){
  # Second argument is used as start date in GA Query
  start <- as.Date(args[2])
}else{
 start <- Sys.Date() - 1
}

if (length(args) == 3){
# Third argument is used as end date in GA Query  
 end <- as.Date(args[3]) 
}else{
 end <- start
}

#*********************************************************************************************************************
# Read arguments and store them in variables -- END
#*********************************************************************************************************************

#*********************************************************************************************************************
# Use arguments values to setup the environment and load initial objects -- START
#*********************************************************************************************************************

# Set working directory to the passed value
setwd(working.directory)

# load Google Analytics ID from Config folder
load("google.analytics.id")

# Load SQL Connection String
load("connection.string")

options(RCurlOptions = list(cainfo = system.file("CurlSSL", "cacert.pem", package = "RCurl")))

# load pre-authentcated Google Analytics object "ga"
load("ga.rga")

# Check if the ga token has expired. If it has then refersh token.
if (ga$isTokenExpired()){
  ga$refreshToken()
}

#*********************************************************************************************************************
# Use arguments values to setup the environment and load initial objects -- END
#*********************************************************************************************************************

# Runs Google Analytics query for the provided id and limits data by date (i.e., start.date = date and end.date = date)
# Wrapped in try/catch to handle scenarios when GA does not return any rows



blog.stats = tryCatch({
 ga$getData(google.analytics.id, 
            start.date = start, 
            end.date = end, 
            metrics = "ga:sessions, ga:users, ga:newUsers, ga:sessionDuration, ga:timeOnPage, ga:pageviews",
            dimensions = "ga:date, ga:pageTitle, ga:medium, ga:hasSocialSourceReferral, ga:source, ga:referralPath",
            sort = "",
            filters = "",
            segment = "",
            start = 1,
            max = 10000)           
 }, warning = function(w) { 
    # print("warning")
    return(NULL)
 }, error = function(e) {
    # print("error")
    return(NULL)
 }, finally = {  
    # print("inside Finally")
}
)

if(length(blog.stats)>0 ){
# Convert Numeric To Integer
  blog.stats[,"sessions"] <- as.integer(blog.stats[,"sessions"])
  blog.stats[,"users"] <- as.integer(blog.stats[,"users"])
  blog.stats[,"newUsers"] <- as.integer(blog.stats[,"newUsers"])
  blog.stats[,"sessionDuration"] <- as.integer(blog.stats[,"sessionDuration"])
  blog.stats[,"timeOnPage"] <- as.integer(blog.stats[,"timeOnPage"])
  blog.stats[,"pageviews"] <- as.integer(blog.stats[,"pageviews"])


# Assign columnnames that match the table's column name in the database
colnames(blog.stats) <- c("Date", "PageTitle", "Medium", "HasSocialSourceReferral", "TrafficSource", "ReferralPath", "Sessions", "Users", "NewUsers", "SessionDurationSeconds", "TimeOnPageSeconds", "PageViews" )

# Open a Connection
myconn <- odbcDriverConnect(connection.string)

# Prepare Delete Query
delete.query <- paste0("DELETE  FROM  GoogleAnalyticsBlogStats WHERE Date >='",start,"' AND Date <='",end,"'")

# Execute Delete Query
sqlQuery(myconn,delete.query,  errors= FALSE)


# Insert rows in the table
sqlSave(myconn, blog.stats, "GoogleAnalyticsBlogStats", safer = FALSE, append = TRUE, rownames = FALSE)

# Close the connection to the database
odbcClose(myconn)
}

步骤 3 --- 从批处理文件执行 Rscript

批处理文件看起来像这样 -

"C:\Program Files\R\R-3.1.0\bin\Rscript.exe" "D:\Social Media Analytics\R Scripts\ExtractGoogleAnalyticsBlogStats.r" "D:\\Social Media Analytics\\R Scripts\\Config"

【讨论】：