How can I order a matrix by another matrix, keeping original numeric value (Using R)

Question

I have searched google a lot, and there were many useful R code solutions in Stackoverflow. I have two matrices, they are index matrix and real return matrix. Their dimensions are same (about 500 by 500 matrix?) They were too big, so I am doing exercise with much smaller example.

> setwd("B:/FE/2013.4.28")
> data <- read.csv("ex.csv")
> data
   a  b   c    d  e  f   g    h     i
1  7  5   2    1 11  4   5   55    22
2  3  1   3    5  2  4   6    8    13
3 90 99 999 9999  2 22 222 2223 10973
4  8  4 988 1004  6 15  12   78    50
> id <- t(apply(data,1,order))
> lapply(1:nrow(id),function(i)data[i,id[i,]])
[[1]]
  d c f b g a  e  i  h
1 1 2 4 5 5 7 11 22 55

[[2]]
  b e a c f d g h  i
2 1 2 3 3 4 5 6 8 13

[[3]]
  e  f  a  b   g   c    h    d     i
3 2 22 90 99 222 999 2223 9999 10973

[[4]]
  b e a  g  f  i  h   c    d
4 4 6 8 12 15 50 78 988 1004

> matrix(names(data)[id],ncol=ncol(data))
     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9]
[1,] "d"  "c"  "f"  "b"  "g"  "a"  "e"  "i"  "h" 
[2,] "b"  "e"  "a"  "c"  "f"  "d"  "g"  "h"  "i" 
[3,] "e"  "f"  "a"  "b"  "g"  "c"  "h"  "d"  "i" 
[4,] "b"  "e"  "a"  "g"  "f"  "i"  "h"  "c"  "d" 
> criteria <- matrix(names(data)[id],ncol=ncol(data))
> data2 <- read.csv("ex2.csv",header=TRUE)
> data2
         a       b        c       d       e        f        g       h         i
1 1.100000 1.13000 0.900000 1.70000 1.54500 1.220000 2.000000 1.40000 1.9800000
2 1.242300 1.64345 1.452500 2.20000 1.43240 0.234423 1.556234 1.32432 1.2342300
3 1.542542 1.35432 1.342523 1.23432 1.43254 1.324320 1.546540 2.43200 0.4321432
4 1.542354 1.65460 0.324130 0.65460 1.23452 1.654325 1.342134 0.34124 1.1000000
> rr <- order(data2,criteria)
> matrix(rr,ncol=ncol(criteria))
     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9]
[1,]   22   16    5   20   28   18    3    6   33
[2,]   12    9   21    2   11   19   17   24   25
[3,]   32    1   34   23    7   10   27    8   14
[4,]   35   36   15   30   29    4   26   13   31

This was my code. I have set 'ex' as index matrix, and ex2 as real return matrix. I wanted to re-order 'ex' ascending row by row (each row means the index(criteria) number in one week)

Then, my definite goal is re-ordering 'matrix ex2' as 'matrix ex'

I had copied and pasted lapply code in stackoverflow. So I could re-order 'matrix ex' like this.

> matrix(names(data)[id],ncol=ncol(data))
     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9]
[1,] "d"  "c"  "f"  "b"  "g"  "a"  "e"  "i"  "h" 
[2,] "b"  "e"  "a"  "c"  "f"  "d"  "g"  "h"  "i" 
[3,] "e"  "f"  "a"  "b"  "g"  "c"  "h"  "d"  "i" 
[4,] "b"  "e"  "a"  "g"  "f"  "i"  "h"  "c"  "d" 

> rr <- order(data2,criteria)
> matrix(rr,ncol=ncol(criteria))
     [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9]
[1,]   22   16    5   20   28   18    3    6   33
[2,]   12    9   21    2   11   19   17   24   25
[3,]   32    1   34   23    7   10   27    8   14
[4,]   35   36   15   30   29    4   26   13   31

data was ex, data2 was ex2. d was header of smallest number in row 1(=1), h was header of largest number in row 1(=55) Therefore, I have succeeded at making ex ordered ascending, row by row. ex was re-ordered row by row. It was presented by headers. I could do this with copy-and paste the code on Stackoverflow.

So, I want to re-order ex2 using the order of ex. Then, I could get only this results. Function 'order' shows me that 22nd number in matrix ex2 was smallest in ex2.

But, I have two problem. First problem is that "I want to re-order ex2 on ex, row by row". My result was ordered, but was not ordered row by row, Second problem was that " I want to know the numeric value of ex2, not the rank", My result was showing me the rank of matrix elements.

How can I get ordered ex2 as the order of ex, using row by row method?

I'm not good at English, I'm sorry for my poor English. Thanks for reading my question!

Answer 1

If you want to order the rows of ex2 according to rows of ex,

idx = t(apply(ex, 1, order))

t(sapply(1:nrow(idx), function(i) ex2[i,][idx[i,]]))