add multiple columns to matrix based on value in existing column

Question

I am looking for a way to add 3 values in 3 different columns to a matrix based on the value in an existing column.

experiment = rbind(1,1,1,2,2,2,3,3,3)
newColumns = matrix(NA,dim(experiment)[1],3) # make 3 columns of length experiment filled with NA
experiment = cbind(experiment,newColumns) # add new columns to the experimental data
experiment = data.frame(experiment)    
experiment[experiment[,1]==1,2:4] = cbind(0,1,2) # add 3 columns at once
experiment$new[experiment[,1]==2] = 5 # add a single column
print(experiment)

  X1 X2 X3 X4 new
1  1  0  0  0  NA
2  1  1  1  1  NA
3  1  2  2  2  NA
4  2 NA NA NA   5
5  2 NA NA NA   5
6  2 NA NA NA   5
7  3 NA NA NA  NA
8  3 NA NA NA  NA
9  3 NA NA NA  NA

this, however, fills the new columns the wrong way. I want column 2 to be all 0's, column 3 to be all 1's and column 4 to be all 3's.

I know I can do it 1 column at a time, but my real dataset is quit large so that isn't my preferred solution. I would like to be able to easily add more columns just by making the range of columns larger and adding values to the 3 values in the example

Answer 1

Instead of this:

experiment[experiment[,1]==1,2:4] = cbind(0,1,2) # add 3 columns at once

Try this:

experiment[experiment[,1] == 1, 2:4] <- rep(c(0:2), each=3)

The problem is that you've provided 3 values (0,1,2) to fill 9 entries. The values are by default filled column-wise . So, the first column is filled with 0, 1, 2 and then the values get recycled . So, it goes again 0,1,2 and 0,1,2. Since you want 0,0,0,1,1,1,2,2,2 , you should explicitly generate using rep(0:2, each=3) (the each does the task of generating the data shown just above).