CODEWITHSUNDEEP
rvectorizationmapply
Sven
Sven

Reputation: 465

Using mapply on a subset in a data.table with a custom function

I have a function with two variables I want to run on a data.frame, but just for a subset. To run the function I use mapply. But the result I get for the subsetted row is the result for excluded row? Maybe you can help? Thanks.

find_s_loss_poisson <- Vectorize(function(lam, target){
  precision <- 5
  if(target > lam){
      return(0)
  }else{
     target <- round(target, precision)
     loss <- lam
     i = 0
     while (round(loss, precision) > target){
           i    <- i + 1
           loss <- loss - (1 - ppois(i - 1, lam, 1))
     }
   return(i)
  }
})


test  <- data.table(W =      c(50, 51),
                    P =      c("A", "B"),
                    Distri = c("NV", "PV"), 
                    DDLT   = c(409, 0.080),
                    Sigma  = c(194, 0.284605),
                    ExpBO  = c(0.2071, 0.104),
                    z      = c(0.48, 9999))

test[Distri == "PV", R := mapply(function(x,y) find_s_loss_poisson(x,y), test$DDLT, test$ExpBO)]
test
find_s_loss_poisson(0.08, 0.1040) # result is 0 and this should be the value in R
find_s_loss_poisson(409, 0.2071) # the result of the 1st line is 449 and this is now in R for the 2nd line?

Upvotes: 1

Views: 206

Answers (1)

MKR
MKR

Reputation: 20085

I dont think mapply is needed in scope of what OP wants to achieve.

Options like:

library(data.table)
test[Distri == "PV", R := find_s_loss_poisson(DDLT,ExpBO)]

OR 

test[, R := ifelse(Distri == "PV", find_s_loss_poisson(DDLT,ExpBO), NA_integer_)]

will be good enough.

Upvotes: 2

Related Questions