R - split a value into the amounts within each layer - r

Question

I'm stumped with this one. I have a starting dataset with 2 columns: an ID and a value.
df <- data.frame(id = c('ABC','XYZ'),
value = c(150, 300))
I then define how I want to 'layer' the values (in this case, I want to split the value into layers of 100).
cut <- seq(0, 300, 100)
So, for the first record of the dataset, the value is 150. I want to split that into the amount within the range 0-100, 100-200 and 200-300.
Starting dataset
id value
ABC 150
XYZ 300
Ending dataset (after defining cut)
id value val_0_100 val_100_200 val_200_300
ABC 150 100 50 0
XYZ 300 100 100 100

Answer 1

You can do it like this:
df <- data.frame(id = c('ABC','XYZ'),
value = c(150, 300))
initial_value = 0
final_value = 300
step = 100
number_of_columns = ceiling(final_value / step)
for (i in 1:number_of_columns){
new_col_name <- paste0("val_", step*(i-1), "_", step*i)
df[,new_col_name] = apply(df["value"] - (step*(i-1)),1, FUN=min,100)
df[,new_col_name] = apply(df[new_col_name],1, FUN=max,0)
}

Answer 2

Here is another way using data.table and dcast
library(data.table)
df <- data.frame(id = c('ABC','XYZ'),
value = c(160, 230))
# Data table
dt <- data.table(df)
# Append Data multiple times based on its value
dt <- dt[rep(seq_len(nrow(dt)), ceiling(dt$value/100)), ]
# cumulative sum to be used in splitting into columns in dcast
dt[, csum := 100]
dt[, csum := cumsum(csum), by = "id"]
# Adding extra column to split into 100s and remainder
dt[, value2 := 100]
dt[csum > value, value2 := value %% 100]
dt[value < 100, value2 := value]
dt_dcast <- dcast(dt, id + value ~ csum, value.var = "value2", fill = 0)
# Rename columns as per the example shown above
colstart <- seq(0, max(dt$csum) - 100, 100)
colend <- seq(100, max(dt$csum), 100)
newname <- c("id", "value", paste0("val_", colstart, "_", colend))
setnames(dt_dcast, names(dt_dcast), newname)