InputData:SQL 預存程式的輸入資料:類別產生器
InputData
:產生 InputData 物件,以擷取輸入參數 (其為資料框架) 的資訊。
在執行指定的查詢時,必須填入資料框架。
此物件是建立預存程序的必要項目,而預存程序中的內嵌 R 函數會採用資料框架輸入參數。
使用方式
InputData(name, defaultQuery = NULL, query = NULL)
引數
name
字元字串,R 函數中會提供給 StoredProcedure 的資料輸入參數名稱。
defaultQuery
指定預設查詢的字元字串,如果在預存程序執行時未提供不同的查詢,預設查詢將會擷取資料。 必須是簡單的 SELECT 查詢。
query
字元字串,指定將用來擷取下一次預存程式中數據的查詢。
值
InputData 物件
範例
## Not run:
# See ?StoredProcedure for creating the "cleandata" table.
# train 1 takes a data frame with clean data and outputs a model
train1 <- function(in_df) {
in_df[,"DayOfWeek"] <- factor(in_df[,"DayOfWeek"], levels=c("Monday","Tuesday","Wednesday","Thursday","Friday","Saturday","Sunday"))
# The model formula
formula <- ArrDelay ~ CRSDepTime + DayOfWeek + CRSDepHour:DayOfWeek
# Train the model
rxSetComputeContext("local")
mm <- rxLinMod(formula, data=in_df)
mm <- rxSerializeModel(mm)
return(list("mm" = mm))
}
# create InpuData Object for an input parameter that is a data frame
# note: if the input parameter is not a data frame use InputParameter object
id <- InputData(name = "in_df",
defaultQuery = paste0("select top 10000 ArrDelay,CRSDepTime,",
"DayOfWeek,CRSDepHour from cleanData"))
# create an OutputParameter object for the variable inside the return list
# note: if that variable is a data frame use OutputData object
out <- OutputParameter("mm", "raw")
# connections string
conStr <- paste0("Driver={ODBC Driver 13 for SQL Server};Server=.;Database=RevoTestDB;",
"Trusted_Connection=Yes;")
# create the stored procedure object
sp_df_op <- StoredProcedure("train1", "spTest1", id, out,
filePath = ".")
# register the stored procedure with the database
registerStoredProcedure(sp_df_op, conStr)
# execute the stored procedure, note: non-data frame variables inside the
# return list are not returned to R. However, if the execution is not successful
# the error will be displayed
model <- executeStoredProcedure(sp_df_op, connectionString = conStr)
# get the linear model
mm <- rxUnserializeModel(model$params$op1)
## End(Not run)