passing rows of a dataframe as a selection parameter to rxdatastep

Question

I have a dataframe like this:

> DataSet_Fehler
    Ohne_Verschiebung    Mit_Verschiebung
1 2016-08-29 19:15:48 2016-08-29 19:19:34
2 2016-08-30 19:38:24 2016-08-30 19:42:18
3 2016-10-28 10:39:24 2016-10-28 10:42:48
4 2016-11-07 19:12:18 2016-11-07 19:15:45

I would like to filter my xdf file based on this dataframe (if I explain it in SQL):

SELECT *
   FROM Myxdf_file
   Where DataSet_Fehler[i,]$Ohne_Verschiebung < Date < DataSet_Fehler[i,]$Mit_Verschiebung

I think transformFunc could be my solution but I am not sure, but I dont know how can I implement it:

Filter_row<-function(DataSet_Fehler)
{
  return(DataSet_Fehler)
}
rxDataStep(inData = MyData,  transformFunc = Filter_row)

How can I do that?

Answer 1

You can pass a function to the rowSelection argument of rxDataStep, that references your data frame:

# dates on which to filter your data
filterDf <- read.csv(text=
"2016-08-29 19:15:48, 2016-08-29 19:19:34
2016-08-30 19:38:24, 2016-08-30 19:42:18
2016-10-28 10:39:24, 2016-10-28 10:42:48
2016-11-07 19:12:18, 2016-11-07 19:15:45
", header=FALSE, colClasses="POSIXct")

# your xdf file
indf <- read.csv(text="dt
2016-08-29 19:16:00
2016-08-29 19:20:00
2016-08-30 19:40:00
2016-09-01 12:00:00
2016-11-07 19:14:00
", colClasses="POSIXct")
inxdf <- rxDataStep(indf, "inxdf.xdf")

rowFilter <- function(x, filterDf)
{
    start <- filterDf[[1]]
    end <- filterDf[[2]]
    vapply(x, function(x) any(start < x & x < end), FUN.VALUE=logical(1))
}

rxDataStep(inxdf,
           rowSelection=fil(dt, filDf),
           transformObjects=list(fil=rowFilter, filDf=filterDf))
#                   dt
#1 2016-08-29 19:16:00
#2 2016-08-30 19:40:00
#3 2016-11-07 19:14:00