Convert long to wide matrix-CodePudding

I would like to perform a following task more efficiently.

n_rows <- 50
n_cols <- 100
data_wide <- matrix(1:(n_rows*n_cols), nrow = n_rows, ncol = n_cols)

count <- 1
data_long <- data.frame()
for (run_row in 1:n_rows) {
  
  for (run_col in 1:n_cols) {
    
    data_long[count, 1:3] <- c(run_row, run_col, data_wide[run_row, run_col])
    
    count <- count  1
  }
  
}

Actually, the number of rows = 980 and columns = 1150 and for loop is slow.

CodePudding user response：

I'm not aware of a built-in solution, but you can construct the columns like so:

f <- function(n_rows, n_cols){
  data_wide <- matrix(1:(n_rows*n_cols), nrow = n_rows, ncol = n_cols)
  
  data.frame(
    row = c(t(row(data_wide))),
    col = c(t(col(data_wide))),
    value = c(t(data_wide))
  )
}

system.time(f(980, 1150))
   user  system elapsed 
   0.04    0.00    0.05

CodePudding user response：

library(tidyverse)

Pivot <- function(rows = 50,cols = 100){
  n_rows <- rows
  n_cols <- cols
  data_wide <- matrix(1:(n_rows*n_cols), nrow = n_rows, ncol = n_cols)
  
# The actual code
  data_wide %>% 
    # tibble is an easier format to work with, for me atleast
    as_tibble() %>% 
    # Since you want each row to have its own unique number
    dplyr::mutate(V1_copy = V1) %>% 
    # The actual conversion from wide to long
    pivot_longer(cols = !matches("V1_copy"),names_to = "V2",values_to = "V3") %>% 
    # Editing to make the output look like yours
    dplyr::mutate(V2 = str_replace(V2,"V","")) %>% 
    dplyr::rename(V1 = V1_copy)}

Output:

> Pivot(50,100)
# A tibble: 5,000 × 3
      V1 V2       V3
   <int> <chr> <int>
 1     1 1         1
 2     1 2        51
 3     1 3       101
 4     1 4       151
 5     1 5       201
 6     1 6       251
 7     1 7       301
 8     1 8       351
 9     1 9       401
10     1 10      451
# … with 4,990 more rows

Speed:

> system.time(Pivot(50,100))
   user  system elapsed 
  0.054   0.033   0.183 
> system.time(Pivot(50,1000))
   user  system elapsed 
  0.122   0.041   0.391 
> system.time(Pivot(50,10000))
   user  system elapsed 
  0.583   0.078   0.748

Speed for the dimensions of your dataset: 980 x 1150

> system.time(Pivot(980,1150))
   user  system elapsed 
  0.661   0.054   0.724