Why aggregate and summarise gives answers in different order?

Question

If I calculate something using aggregate function or using summarise in dplyr package why those gives answers different order?

Example:

a <- aggregate(hp~mpg+cyl+gear, mtcars, FUN = sum)

gives me

 mpg cyl gear  hp
1  21.5   4    3  97
2  18.1   6    3 105
3  21.4   6    3 110
4  10.4   8    3 420
5  13.3   8    3 245

and

b <- mtcars %>%
     group_by(mpg, cyl, gear) %>%
     summarise(hp = sum(hp))

gives me

    mpg   cyl  gear    hp
   <dbl> <dbl> <dbl> <dbl>
 1  10.4     8     3   420
 2  13.3     8     3   245
 3  14.3     8     3   245
 4  14.7     8     3   230
 5  15       8     5   335

Why order is not the same?

Answer 1

As mentioned by @zx8754, tidyverse operations will re-order the rows. No guarantee that you will get a certain row order. https://github.com/tidyverse/dplyr/issues/2192#issuecomment-281655703

Looking a bit closely, I see that aggregate sorted by gear, cyl, then mpg. So the following tidyverse code will provide the same row order as aggregate(hp~mpg+cyl+gear, mtcars, FUN = sum) :

library(tidyverse)
mtcars %>% group_by(gear, cyl, mpg) %>% summarise(hp = sum(hp)) %>% head()
#> # A tibble: 6 x 4
#> # Groups:   gear, cyl [3]
#>    gear   cyl   mpg    hp
#>   <dbl> <dbl> <dbl> <dbl>
#> 1     3     4  21.5    97
#> 2     3     6  18.1   105
#> 3     3     6  21.4   110
#> 4     3     8  10.4   420
#> 5     3     8  13.3   245
#> 6     3     8  14.3   245

^{Created on 2019-02-27 by the reprex package (v0.2.1)}

and to get the same row order as mtcars %>% group_by(mpg, cyl, gear) %>% summarise(hp = sum(hp)):

library(tidyverse)
aggregate(hp~gear+cyl+mpg, mtcars, FUN = sum) %>% head()
#>   gear cyl  mpg  hp
#> 1    3   8 10.4 420
#> 2    3   8 13.3 245
#> 3    3   8 14.3 245
#> 4    3   8 14.7 230
#> 5    5   8 15.0 335
#> 6    3   8 15.2 330

^{Created on 2019-02-27 by the reprex package (v0.2.1)}