Microsoft R Server 逐行插入

Microsoft R Server Row by Row Insert

我有一个通常写入平面文件的 for 循环。这样,如果有任何问题,我可以从我离开的地方开始。我想在执行我的 R 代码的 SQL Server 2016 存储过程中直接使用新的 RevoScaleR 函数将此过程转换为从 SQL table 读取和写入。

这是一个简单的 SPROC:

USE [master]
GO

/****** Object:  StoredProcedure [dbo].[Rscript_geocodeUSACities_TEST]    Script Date: 8/8/2017 11:40:40 AM ******/
SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO




CREATE PROCEDURE [dbo].[Rscript_geocodeUSACities_TEST]
    @usrOutputFilePath varchar(150)
    ,@usrOutputFileName varchar(150)

AS
BEGIN

    SET NOCOUNT ON;

DECLARE @rScript nvarchar(max) = N'

#### USER INPUTS ####

usrOutputFile <- "' + @usrOutputFilePath + @usrOutputFileName + '"


#### ESTABLISH ENVIRONMENT ####

library(data.table)
library(foreach)
library(XML)
library(RCurl)
library(RJSONIO)

##turn off scientific notation
options(scipen=999)

##establish compute context
sqlServerConnString <- "Server=.;Database=External;Trusted_Connection=true"
sqlServerCC <- RxInSqlServer(connectionString=sqlServerConnString)
rxSetComputeContext(sqlServerCC)
print(rxGetComputeContext())


#### GEOCODE ####

print(dfInputData)
rxDataStep(data=dfInputData,outFile=imp.USA_Cities_Map,append="rows")

'

EXECUTE  sp_execute_external_script
                @language = N'R'
              , @script = @rScript
              ,@input_data_1 =N'select 5 as test_insert'
            ,@input_data_1_name =N'dfInputData'
              ;

END

错误输出:

Error in rxDataStep(data = dfInputData, outFile = imp.USA_Cities_Map,  : 
  object 'imp.USA_Cities_Map' not found

给你。您不需要将计算上下文设置为 SQL 服务器。但是您必须向本地用户授予登录权限 运行 R 外部进程。它们都被添加到名为 SqlRUserGroup 的本地组中,您只需将 'dbrownebook' 替换为您的服务器名称即可。

请注意,您没有为 sqlrusergroup 添加数据库用户,而只是添加了登录名。 SQL R 服务将模拟调用 sp_execute_external_script 的用户。解释如下:https://docs.microsoft.com/en-us/sql/advanced-analytics/r/security-considerations-for-the-r-runtime-in-sql-server

use master
go

create login [dbrownebook\sqlrusergroup] from windows

create database [External]

go

use [External]
go

create schema imp
go
create table imp.USA_Cities_Map(test_insert int)
go


/****** Object:  StoredProcedure [dbo].[Rscript_geocodeUSACities_TEST]    Script Date: 8/8/2017 11:40:40 AM ******/
SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO




CREATE OR ALTER PROCEDURE [dbo].[Rscript_geocodeUSACities_TEST]
    @usrOutputFilePath varchar(150)
    ,@usrOutputFileName varchar(150)

AS
BEGIN

    SET NOCOUNT ON;

DECLARE @rScript nvarchar(max) = N'

sqlServerConnString <- "Server=.;Database=External;Trusted_Connection=true"
sqlTable <- RxSqlServerData(table = "imp.USA_Cities_Map", connectionString = sqlServerConnString)

rxDataStep(data=dfInputData,outFile=sqlTable,append="rows")
rxDataStep(data=dfInputData,outFile=sqlTable,append="rows")
rxDataStep(data=dfInputData,outFile=sqlTable,append="rows")

'

EXECUTE  sp_execute_external_script
                @language = N'R'
              , @script = @rScript
              ,@input_data_1 =N'select 5 as test_insert'
            ,@input_data_1_name =N'dfInputData'
              ;

END

GO

exec [Rscript_geocodeUSACities_TEST] '',''

go
select * from imp.USA_Cities_Map