Error in get(as.character(FUN), mode = “function”, envir = envir) :

Question

I am trying apply this function on a data frame to create a new feature but I keep getting this error:

Error in get(as.character(FUN), mode = "function", envir = envir) : object 'INSTALLMENT' of mode 'function' was not found

I have also tried apply(df, 2, purchase) but then I get this error:

Error: $ operator is invalid for atomic vectors

The code is :

purchase = function(DataFrame){

  if((DataFrame$ONEOFF_PURCHASES == 0) && (DataFrame$INSTALLMENTS_PURCHASES == 0))
    return('NONE')

  if((DataFrame$ONEOFF_PURCHASES > 0) && (DataFrame$INSTALLMENTS_PURCHASES > 0))
    return('BOTH_ONEOFF_INSTALLMENT')

  if((DataFrame$ONEOFF_PURCHASES > 0) && (DataFrame$INSTALLMENTS_PURCHASES == 0))
    return('ONE_OFF')

  if((DataFrame$ONEOFF_PURCHASES == 0) && (DataFrame$INSTALLMENTS_PURCHASES > 0))
    return('INSTALLMENT')
}

df$PURCHASE_TYPE = apply(df, 2, purchase(df))

Answer 1

It looks like a ifelse() case. if() ... else ... can only treat a single boolean. To make your function work, you need to vectorize it by mapply() or Vectorize() .

Example Data

set.seed(1)
df <- data.frame(ONEOFF_PURCHASES = sample(0:1, 5, T), INSTALLMENTS_PURCHASES = sample(0:1, 5, T))

#   ONEOFF_PURCHASES INSTALLMENTS_PURCHASES
# 1                0                      0
# 2                1                      0
# 3                0                      0
# 4                0                      1
# 5                1                      1

Function

purchase <- function(x, y){
  if((x == 0) && (y == 0))
    return('NONE')
  else if((x > 0) && (y > 0))
    return('BOTH_ONEOFF_INSTALLMENT')
  else if((x > 0) && (y == 0))
    return('ONE_OFF')
  else if((x == 0) && (y > 0))
    return('INSTALLMENT')
  else
    return('OTHERS')
}

Vectorization

mapply(purchase, df$ONEOFF_PURCHASES, df$INSTALLMENTS_PURCHASES)
# [1] "NONE"  "ONE_OFF"  "NONE"  "INSTALLMENT"  "BOTH_ONEOFF_INSTALLMENT"

Vectorize(purchase)(df$ONEOFF_PURCHASES, df$INSTALLMENTS_PURCHASES)
# [1] "NONE"  "ONE_OFF"  "NONE"  "INSTALLMENT"  "BOTH_ONEOFF_INSTALLMENT"

---

Actually, we don't use the approach above in this issue. We'll use ifelse() or dplyr::case_when() .

library(dplyr)

df %>%
  mutate(PURCHASE_TYPE = case_when(
    (ONEOFF_PURCHASES == 0) & (INSTALLMENTS_PURCHASES == 0) ~ 'NONE',
    (ONEOFF_PURCHASES > 0) & (INSTALLMENTS_PURCHASES > 0) ~ 'BOTH_ONEOFF_INSTALLMENT',
    (ONEOFF_PURCHASES > 0) & (INSTALLMENTS_PURCHASES == 0) ~ 'ONE_OFF',
    (ONEOFF_PURCHASES == 0) & (INSTALLMENTS_PURCHASES > 0) ~ 'INSTALLMENT',
    TRUE ~ 'OTHERS'
  ))

#   ONEOFF_PURCHASES INSTALLMENTS_PURCHASES           PURCHASE_TYPE
# 1                0                      0                    NONE
# 2                1                      0                 ONE_OFF
# 3                0                      0                    NONE
# 4                0                      1             INSTALLMENT
# 5                1                      1 BOTH_ONEOFF_INSTALLMENT