CODEWITHSUNDEEP
rdataframedata.table
Nicso
Nicso

Reputation: 143

Alternative in data.table

This is what I usually do every time I want to create a new column in a data.frame, the new column being the result of a function applied to different subsets of my data.frame.

What do you think is the best way to get the same result using data.table package in R?

Cheers!

> class(DF)
[1] "data.frame"
> names(DF)
[1] "sp"  "X1"  "X2"

paramsVal <- c(0.32, 0.23, 8.28, 8.37)


DF <- split(DF, DF$sp)

DF <- lapply(seq_along(DF), function(X){

    Data <- DF[[X]]

    if(unique(X$sp) == "SP1"){
    Data$Pred <- fakeFunction(Data = Data,
                              param1 = paramsVal[1],
                              param2 = paramsVal[3])
    }else{
        Data$Pred <- fakeFunction(Data = Data,
                                  param1 = paramsVal[2],
                                  param2 = paramsVal[4])
    }

    return(Data)
})

DF <- do.call("rbind", DF)

names(DF)
[1] "sp"  "X1"  "X2"  "Pred"

Upvotes: 0

Views: 259

Answers (1)

Gregor Thomas
Gregor Thomas

Reputation: 145975

With data.table, I would do this:

DT = as.data.table(DF)
DT[sp == "SP1", Pred := fakeFunction(Data = .SD, param1 = paramsVal[1], param2 = paramsVal[3])]
DT[sp != "SP1", Pred := fakeFunction(Data = .SD, param1 = paramsVal[2], param2 = paramsVal[4])]

I think this should work, but I can't test without a reproducible example. If you need more assistance, please provide (a) a copy/pasteable sample of the data (just a couple rows each of SP1 and not SP1 - use dput() for reproducibility), and (b) a stand-in for fakeFunction, parmsVal, and anything else needed for the example to run.

Upvotes: 1

Related Questions