Alternate, interweave or interlace two matrices

Question

I've got two matrices which I want to interweave/interlace/stack alternatively on top of each other/rbind in an alternating manner.

ranks=1:3
names=c("Karl", "Klaus", "Mary")
x <- cbind(ranks, names)

universities=c("Cape Town", "London", "Berlin")
y <- cbind(rep("", 3), universities)

After one row of x, I want one row of y:

     [,1] [,2]       
[1,] "1"  "Karl"     
[2,] ""   "Cape Town"
[3,] "2"  "Klaus"      
[4,] ""   "London"   
[5,] "3"  "Mary"   
[6,] ""   "Berlin"

I tried matrix(rbind(x, y), ncol=2) (which seems to solve the problem if I had two strings) without effect

Transposing! Great idea. Want upgrade your comment to an answer? — MERose, Dec 29 '15 at 00:56
Not to state the obvious, but why would you do this? Column 2 would now have two different types of data with no indication of what type of variable it is--only the assumption that the pattern is always going to be "name, university". Do we have to start with "x" and "y"? — A5C1D2H2I1M1N2O1R2T1, Dec 29 '15 at 02:31
I am building a Latex-Table - that's why I need this. Therefore the order matters. — MERose, Dec 29 '15 at 10:17

score 1 · Accepted Answer · answered Dec 29 '15 at 02:40

Here are two alternatives.

First, assuming we have to start with "x" and "y", you can try interleave from the "gdata" package:

library(gdata)
interleave(x, y)
#      ranks names      
# [1,] "1"   "Karl"     
# [2,] ""    "Cape Town"
# [3,] "2"   "Klaus"    
# [4,] ""    "London"   
# [5,] "3"   "Mary"     
# [6,] ""    "Berlin"

Second, assuming we can start with "ranks", "names", and "universities", you can use base R, like this:

cbind(c(t(cbind(ranks, ""))), c(t(cbind(names, universities))))
#      [,1] [,2]       
# [1,] "1"  "Karl"     
# [2,] ""   "Cape Town"
# [3,] "2"  "Klaus"    
# [4,] ""   "London"   
# [5,] "3"  "Mary"     
# [6,] ""   "Berlin"

A better alternative, however, would be to use something like melt (from "reshape2" or "data.table"). This would allow you to add another variable that indicates what type of measurement a value represents.

library(data.table)
melt(data.table(ranks, names, universities), "ranks")
#    ranks     variable     value
# 1:     1        names      Karl
# 2:     2        names     Klaus
# 3:     3        names      Mary
# 4:     1 universities Cape Town
# 5:     2 universities    London
# 6:     3 universities    Berlin

Or, to match your desired ordering:

library(data.table)
setorder(melt(data.table(ranks, names, universities), "ranks"), ranks)[]
#    ranks     variable     value
# 1:     1        names      Karl
# 2:     1 universities Cape Town
# 3:     2        names     Klaus
# 4:     2 universities    London
# 5:     3        names      Mary
# 6:     3 universities    Berlin

Alternate, interweave or interlace two matrices

1 Answers1