Get full dataset

getFullDataset.R

# Define the SQL Connection string and the query
sqlServerConnString <- "SERVER=winrserver;DATABASE=AdventureWorksDW2017;UID=ruser;PWD=rpass"
qry <- "select  
       [Age]
      ,[MaritalStatus]
      ,[Gender]
      ,[YearlyIncome]
      ,[TotalChildren]
      ,[NumberChildrenAtHome]
      ,[Education]
      ,[Occupation]
      ,[IsHomeOwner]
      ,[NumberCarsOwned]
      ,[YearsSinceFirstPurchase]
      ,[CommuteDistance]
      ,[CountryRegionCode]
      ,[StateProvinceCode]
      ,[IsCardUser]
from dbo.vwDimCustomer"

# Get the results into a dataset
sqlData <- RxSqlServerData(sqlQuery = qry, connectionString = sqlServerConnString,stringsAsFactors = TRUE)
ds <- rxDataStep(sqlData,reportProgress=0)

# Factorize some continuous values in order to avoid nonsensical value splits such as <2.5 children and 1.5 cars
ds$TotalChildren <- as.factor(ds$TotalChildren)
ds$NumberChildrenAtHome <- as.factor(ds$NumberChildrenAtHome)
ds$NumberCarsOwned <- as.factor(ds$NumberCarsOwned)
ds$YearsSinceFirstPurchase <- as.factor(ds$YearsSinceFirstPurchase)