How to "unmelt" the dataframe into matrix

July 19, 2022

I have a dataframe

sources <- c("a", "b", "c", "a", "a", "b")
targets <- c("f", "v", "w", "a", "b", "c")
values <- c(12, 34, 56, 3, 76, 35)
df <- data.frame(sources, targets, values)

  sources targets values
1       a       f     12
2       b       v     34
3       c       w     56
4       a       a      3
5       a       b     76
6       b       c     35

How to reshape it to get a matrix with sources as row names, targets as colnames and corresponding values in intersecting cells? In this case, I should have a 3 by 6 matrix. something like this:

    f  v  w  a  b  c
a  12  0  0  3  76 0
b  0  34  0  0  0 35 
c  0  0   56 0  0  0

>Solution :

You may try

library(tidyverse)

df <- data.frame(sources, targets, values)

df %>%
  pivot_wider(names_from = targets, values_from = values) %>%
  replace(is.na(.), 0) %>%
  column_to_rownames(var = "sources")


   f  v  w a  b  c
a 12  0  0 3 76  0
b  0 34  0 0  0 35
c  0  0 56 0  0  0