R - recursion over columns of data.table

Question

I am given data-table, where the (i+1)-th column depends on the previous one and needs to be calculated recursively. The header of data-table consists of a sequence starting from 0.`

   0  1  2  3
1: 1 NA NA NA
2: 2 NA NA NA
3: 3 NA NA NA

Accessing the columns by the index, e.g. dt[,..2] makes the code IMHO the most readable. Yet, this apparently cannot be used when trying to assign:

library(data.table)
dt <- data.table("0"=c(1,2,3),"1"=c(NA,NA,NA),"2"=c(NA,NA,NA),"3"=c(NA,NA,NA))
x <- c(0.01, 0.02, 0.015)

for (u in 1:3){
               v<- u+1
               dt[,..v] <- dt[,..u]*(1+x[u])
 }

This yields the following error:

Error in `[<-.data.table`(`*tmp*`, , ..v, value = list(`0` = c(1.01, 2.02,  : object '..v' not found

Update: Thanks @IceCreamToucan for the answer. However, I have just posted a simple example to illustrate my general issue. Due to the higher complexity of my actual code, I probably will have to stick to a for-loop. So I look to a solution, to recursively access and assign the columns.

IceCreamToucan · Accepted Answer · 2019-06-06 17:18:51Z

2

You can do this with lapply and cumprod. I renamed the varibles because I'm not sure how to deal with numeric column names. Also see comments for an outer and cumprod option.

setnames(dt, names(dt), paste0('v', names(dt)))
dt[, names(dt)[-1] := lapply(cumprod(1 + x), '*', v0)][]

#    v0   v1     v2       v3
# 1:  1 1.01 1.0302 1.045653
# 2:  2 2.02 2.0604 2.091306
# 3:  3 3.03 3.0906 3.136959

You can also do it this way

for (u in 0:2){
  v <- u+1
  dt[, as.character(v) := get(as.character(u))*(1 + x[u + 1])] 
}

dt[]

#    0    1      2        3
# 1: 1 1.01 1.0302 1.045653
# 2: 2 2.02 2.0604 2.091306
# 3: 3 3.03 3.0906 3.136959

edited Jun 6, 2019 at 17:18

answered Jun 6, 2019 at 17:10

IceCreamToucan

28.8k2 gold badges27 silver badges48 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Frank Over a year ago

Or use a matrix dt[, outer(`0`, cumprod(1+c(0,x)))] (fwiw, numeric and other exotic col names can be accessed with a backtick, but I'd recommend renaming as you do here anyways)

Collectives™ on Stack Overflow

R - recursion over columns of data.table

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related