R：根据 dplyr 的列值打破 data.frame

Question

I have this data.frame

MWE <- data.frame(x = c("a", "a", "a", "b", "b", "b"), y = c(1,2,3,4,5,6))

and what I want to obtain is this data.frame

data.frame(a = c(1,2,3), b = c(4,5,6)) 

Actually, what I originally want is to sum the 2 vectors a and b (well, I have in reality many more vectors, but it is easier to explain with only 2), so that's why I thought about this transformation. I can do a rowSums then, or something equivalent.

I tried to use pivot_wider from tidyr but I had an error.

Any idea of how to do this with dplyr or tidyr ?

Answer 1

Continuing from @Mr.Flick's attempt in tidyverse you could create an id column and grouped on that id column calculate the sum like

library(dplyr)

MWE %>%
  group_by(x) %>%
  mutate(row = row_number()) %>%
  group_by(row) %>%
  mutate(total_sum = sum(y)) %>%
  tidyr::pivot_wider(names_from = x, values_from = y) %>%
  ungroup() %>%
  select(-row)

# A tibble: 3 x 3
#  total_sum     a     b
#      <dbl> <dbl> <dbl>
#1         5     1     4
#2         7     2     5
#3         9     3     6

Answer 2

We can use unstack from base R

unstack(MWE, y ~ x)
#  a b
#1 1 4
#2 2 5
#3 3 6

---

Or using rowid from data.table with pivot_wider from tidyr

library(dplyr)
library(data.table)
library(tidyr)
MWE %>% 
   mutate(rn = rowid(x)) %>% 
   pivot_wider(names_from = x, values_from = y) %>%
   select(-rn)
# A tibble: 3 x 2
#      a     b
#  <dbl> <dbl>
#1     1     4
#2     2     5
#3     3     6

Answer 3

Using base R:

data.frame(with(MWE, split(y, x)))

  a b
1 1 4
2 2 5
3 3 6