Why is the result of `==` of class integer when used inside ave()?

Question

This post is based on this question: Running ifelse statement by a variable

Given some data

  col1 col2
1    B   10
2    B    7
3    C    7
4    D    1
5    A    3
6    D    2

to which we want to add a new column that indicates whether the value of 'col2' is the maximum of 'col2' grouped by 'col1'.

In base R one can use transform and ave

transform(df, new_col = ave(col2, col1, FUN = function(x) x == max(x)))
  col1 col2 new_col
1    B   10       1
2    B    7       0
3    C    7       1
4    D    1       0
5    A    3       1
6    D    2       1

My question is why 'new_col' is of type integer and not logical?

@RichScriven pointed out in the comments that that's because 'col2' is already numeric but I don't understand that.

Thanks!

data

df <- structure(list(col1 = c("B", "B", "C", "D", "A", "D"), col2 = c(10L, 
7L, 7L, 1L, 3L, 2L)), .Names = c("col1", "col2"), row.names = c(NA, 
-6L), class = "data.frame")

Answer 1

ave will return an object of the same class as x if it can be coerced

with(df, ave(col2, col1, FUN = function(x) T))
# [1] 1 1 1 1 1 1
with(df, ave(col2, col1, FUN = function(x) 'a'))
# [1] "a" "a" "a" "a" "a" "a"

df$col2 <- as.logical(df$col2)
with(df, ave(col2, col1, FUN = function(x) T))
# [1] TRUE TRUE TRUE TRUE TRUE TRUE



df$col2 <- as.factor(df$col2)
with(df, ave(col2, col1, FUN = function(x) T))
# [1] TRUE TRUE TRUE TRUE TRUE TRUE
# Levels: TRUE

Edit: It appears that it will coerce to a "higher" class but not a "lower" class. Where "higher" and "lower" is determined by the ordering given in ?c (below).

df$col2 <- as.logical(df$col2)
with(df, ave(col2, col1, FUN = function(x) 1))
# [1] 1 1 1 1 1 1

NULL < raw < logical < integer < double < complex < character < list < expression