R中的默认列数和行数

Question

In R, I can create a matrix with matrix() , which has the function definition:

function (data = NA, nrow = 1, ncol = 1, byrow = FALSE, dimnames = NULL)
{ ... }

This suggests that the default value for number of rows ( nrow ) and columns ( ncol ) are 1.

So why does the following break? Why does specifying the same default values of 1 result in a different matrix?

> matrix(1:9)
      [,1]
 [1,]    1
 [2,]    2
 [3,]    3
 [4,]    4
 [5,]    5
 [6,]    6
 [7,]    7
 [8,]    8
 [9,]    9

> matrix(1:9, ncol=1, nrow=1)
     [,1]
[1,]    1

Answer 1

The user-facing base::matrix calls an internal function :

matrix
# function (data = NA, nrow = 1, ncol = 1, byrow = FALSE, dimnames = NULL) 
# {
#     if (is.object(data) || !is.atomic(data)) 
#         data <- as.vector(data)
#     .Internal(matrix(data, nrow, ncol, byrow, dimnames, missing(nrow), 
#         missing(ncol)))
# }
# <bytecode: 0x000000000b802e88>
# <environment: namespace:base>

In particular, note that the last two arguments passed to the internal function are flags indicating if the row and / or column dimensions were specified: missing(nrow), missing(ncol) . The fact that these parameters have default values does not preclude them from being "missing". For example,

f <- function(x = 1, y = 2) {
    cat(sprintf(
        "missing(x): %s\nmissing(y): %s\n", 
        missing(x), 
        missing(y)
    ))
}

f()
# missing(x): TRUE
# missing(y): TRUE

---

Since the "missing-ness" of these arguments is not affected by the fact that they have default values, that logic can be handled independently, as is done here .