一次有条件地将值分配给两列 R

Question

Problem:

Is there a way to write a single command to assign different values to two columns of a row that meets one condition ?

Context:

I have to conditionally assign a value to two columns: Quantity and Price . I want to do this using the data.table library in R.

My dataset looks like this:

> example
tariff  expenditure  
     1           50
     2           70
     3           50

Each tariff has different prices. Thus, a single expenditure can result in different consumption quantities.

Suppose the prices of tariffs 1, 2 and 3 are 10, 20 and 30 respectively. I want to do the following in a single expression:

1. If tariff is 1, then Consumption is expenditure/10 and Price is 10 .
2. If tariff is 2, then Consumption is expenditure/20 and price is 20 .
3. If tariff is 3, then Consumption is expenditure/30 and Price is 30 .

In data.table, this can be done by slicing example as follows:

example[tariff == 1, c("Consumption", "Price") := list(expenditure / 10, 10)]

In the real dataset, Consumption and Price depend on the value of many other columns besides tariff . If I proceed as shown above, I will end up with about 100 different boolean slices. I'd much rather do this using ifelse .

The following code failed:

example[, c("Consumption", "Price") := ifelse(tariff == 1, list(expenditure/10, 10),
                                              ifelse(tariff == 2, list(expenditure/20, 20),
                                                     list(expenditure/30, 30)))]

Is there a way to do this?

Answer 1

Here's a clever way to do it (since your numbers are so simple in this case)!

#first make this a dataframe so tidyverse functions can interpret it
d <- as.data.frame(list(tariff = c(1, 2, 3), expenditure = c(50, 70, 50)))
#mutate can create new variables, and your transformation is the same in each condition
d %>% mutate("Consumption" = expenditure/(tariff*10), "Price" = 10*tariff)

Answer 2

A possible approach is to store your functions in a column of a transformation table (eg tfn here). Join this table with your dataset and apply the function to relevant column.

library(data.table)
#sample transformation
(tfn <- data.table(ID=LETTERS[1L:3L], 
    tariff=1L:3L, 
    consumpF=list(function(x) x/10, function(x) x/20, function(x) x/30), 
    priceF=list(function(x) (x-1)*10, function(x) x*10, function(x) x*20)))

#sample dataset
(ds <- data.table(ID=LETTERS[1L:3L], 
    tariff=1L:3L, 
    expenditure=seq(10, 30, 10)))

#join and apply function on values
ds[tfn, on=.(ID, tariff), `:=` (
    Consumption = mapply(function(f,x) f(x), consumpF, expenditure),
    Price = mapply(function(f,x) f(x), priceF, tariff)
)]