0

I would like to make subgroups for a matrix， and each subgroup contains different amount of columns. For example, there is a matrix with 1000 rows and 420 columns, and I would like to split into 35 sub-matrix in order, like first one contains the first 12 cols, and the second contains the first 24 cols, the third one contains the first 36 cols so on?

Amanda
  • 63
  • 1
  • 7
  • 1
    Hello. Please include a reproducible example along with expected output and code you have tried so far. Additionally, [look at this link](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) for further info. – Sotos May 10 '16 at 09:02
  • Did you calculated the need of RAM? Do you need all submatrices at once? – jogo May 10 '16 at 09:19
  • Possible duplicate of [split big matrix in subsets,R](http://stackoverflow.com/questions/37113620/split-big-matrix-in-subsets-r) – jogo May 10 '16 at 10:02

1 Answers1

0

If M is your matrix with 1000 rows and 420 columns
this will give you a list of the desired matrices:

lapply(seq(12, 420, by=12), function(x) M[, 1:x])

This will need a lot of RAM. If you don't need all submatrices at once you can do a loop:

for (x in seq(12, 420, by=12)) {
  subM <- M[, 1:x]
  # .... use subM
}
jogo
  • 12,469
  • 11
  • 37
  • 42
  • Thanks Jogo! Yes, I will need to have all matrices at once, possible to achieve that? – Amanda May 10 '16 at 09:33
  • 1
    In which form you need the matrices? Please edit your question http://stackoverflow.com/posts/37133542/edit The `lapply()` gives you a list of the matrices. You can thank by accepting the answer. – jogo May 10 '16 at 09:38