Updating a data.table in R

Question

(Edited)

I am using the following code to create two columns in a data.table and update them with some numbers:

T <- data.table(Init_1 = rep(0, 100), Init_2 = rep(0, 100))

for (i in 1:100){
  T[, Init_1 := i]
  T[, Init_2 := 2*i]
}

I expected that this code would add two columns to the data.table T (Init_1 and Init_2) and fill them with numbers : (1:100) and (2,4,...200) respectively.

However, the code returns constant values:

     > T
   Init_1 Init_2
  1:    100    200
  2:    100    200
  3:    100    200
  4:    100    200
  5:    100    200
  6:    100    200
  7:    100    200
  8:    100    200
.................

Could you explain why my code is not working as expected and how it could be fixed?

Your advice will be appreciated.

---

Edit:

In relation to answer 2, eventually I want to use a function inside the for loop. More specifically:

# A FUNCTION THAT RETURNS THE TRANSITION PROBABILITIES AFTER N STEPS IN A MARKOV CHAIN
#-------------------------------------------------------------------------------------
R <- function(P, n){
  if (n==1) return(P)
  R(P, n-1) %*% P

}

# A ONE-STEP PROBABILITY MATRIX
#---------------------------------------------------------------------------------------
P = matrix(c(0.6, 0.1, 0.3, 0.2, 0.7, 0.1, 0.3, 0.3, 0.4), nrow = 3, byrow = TRUE)

# EXAMINING THE CONVERGENCE PROCESS OF THE PROBABILITIES OVER TIME
#########################################################################
T <- data.table(Init_1 = rep(0, 100), Init_2 = rep(0, 100))

for (i in 1:100){
  T[, Init_1 := R(P, i)[1,1]]
  T[, Init_2 := R(P, i)[2,1]]
}

or

   for (i in 1:100){
      T[, ':=' (Init_1 = R(P, i)[1,1],
                Init_2 = R(P, i)[2,1]) ]
    }

Answer 1

I'm no data.table expert. But I know it throws helpful error messages. If you e.g. create an empty data.table and try to use := to add columns, it says

T <- data.table()
T[,a:=1]
# Error in `[.data.table`(T, , `:=`(a, 1)) : 
#   Cannot use := to add columns to a null data.table (no columns), currently. 
#   You can use := to add (empty) columns to a 0-row data.table (1 or more empty columns), 
#   though.

Your problem might be related. Because data.table(numeric()) or rather T <- data.table(numeric(length = 0)) creates a a 0-row data.table. The empty column gets named V1 by default. Here you could use := to add empty columns. However, that's not what you want.

Instead you could do

T <- data.table(numeric(0))
for (i in 1:5){
  T <- T[, .(
    Init_1=if (exists("Init_1")) c(Init_1, i) else i, 
    Init_2=if (exists("Init_2")) c(Init_2, 2*i) else  2*i )]
}
T
#    Init_1 Init_2
# 1:      1      2
# 2:      2      4
# 3:      3      6
# 4:      4      8
# 5:      5     10

Although that's pretty ugly und probably super unefficient.

Answer 2

First, you should not define a function with name as T. T is reserved for TRUE in logic. Also, it is not recommended to use i for iteration since it is also used for complex number, for example

> (2i)^2
[1] -4+0i

Third, iteration is slow in R. We should avoid to use iteration whenever possible.

Here are the simple codes to generate such matrix. Hope this helps.

T.data <- matrix(NA,nrow=100,ncol=2);
T.data[,1] <- 1:100;
T.data[,2] <- 2*T.data[,1]