如何将连续变量重新编码为范围

Question

I need to recode a continuous variable into categories, usually i use the "cut" Function, but in the cut function i need to specify the breaks. i am looking for a way to have a different set of breaks depending on other categorical variables in my data frame.

the variable in my example is Cost and the "breaks" are in the second table "cost.range", i have a different set of Breaks for each "Region" and each "Category"

Example :

Region    Product     Category Cost
Country A Product 1     CAT A  731
Country B Product 1     CAT A  659
Country C Product 1     CAT A  385
Country D Product 1     CAT A  763
Country A Product 2     CAT A  701
Country B Product 2     CAT A  759
Country C Product 2     CAT A  580
Country D Product 2     CAT A  147
Country A Product 3     CAT B  645
Country B Product 3     CAT B  657
Country C Product 3     CAT B  424


Region       Category  Cost.Range      Range
Country A      CAT A         10          R1
Country A      CAT A         50          R2
Country A      CAT A        200          R3
Country A      CAT A       1000          R4
Country A      CAT B         20          R1
Country A      CAT B        100          R2
Country A      CAT B        400          R3
Country A      CAT B       1500          R4

code to generate the example :

Region <- c("Country A","Country B","Country C","Country D","Country A","Country B","Country C","Country D","Country A","Country B","Country C","Country D","Country A","Country B","Country C","Country D")
Product <- c("Product 1","Product 1","Product 1","Product 1","Product 2","Product 2","Product 2","Product 2","Product 3","Product 3","Product 3","Product 3","Product 4","Product 4","Product 4","Product 4")
Category <- c("CAT A","CAT A","CAT A","CAT A","CAT A","CAT A","CAT A","CAT A","CAT B","CAT B","CAT B","CAT B","CAT B","CAT B","CAT B","CAT B")
Cost <- c(731,659,385,763,701,759,580,147,645,657,424,34,850,463,160,550)

Table1 <- data.frame(Region, Product, Category, Cost)

Region <- c("Country A","Country A","Country A","Country A","Country A","Country A","Country A","Country A")
Category <- c("CAT A","CAT A","CAT A","CAT A","CAT B","CAT B","CAT B","CAT B")
Cost.range <- c(10,50,200,1000,20,100,400,1500)
Range <- c("R1","R1","R3","R4","R1","R2","R3","R4")

Table2 <- data.frame(Region, Category, Cost.range, Range)

Answer 1

This is not the most elegant solution (and I'd be interested to see a better method) but it should achieve the result you're looking for.

The select() and distinct() functions from the dplyr package find the possible combinations of Region and Category . These combinations are used to subset the two tables and apply the cut() function to each subset.

library('dplyr')
library('data.table')

dt1 <- data.table(Table1)
dt2 <- data.table(Table2)

t2d <- Table2 %>% select(Region, Category) %>% distinct

for(i in 1:nrow(t2d)){
  dt2_range_subset <- dt2[Region == as.character(t2d$Region[i]) 
                          & Category == t2d$Category[i], Cost.range]
  dt1[Region == as.character(t2d$Region[i]) & Category == t2d$Category[i],
      Cost_factor := cut(as.matrix(Cost), dt2_range_subset)]
}