This is actually a bit trickier than you'd think. Since a list can actually (with some effort) contain NULL elements, it might not be enough to check is.null(foo$a)
. A more stringent test might be to check that the name is actually defined in the list:
foo <- list(a=42, b=NULL)
foo
is.null(foo[["a"]]) # FALSE
is.null(foo[["b"]]) # TRUE, but the element "exists"...
is.null(foo[["c"]]) # TRUE
"a" %in% names(foo) # TRUE
"b" %in% names(foo) # TRUE
"c" %in% names(foo) # FALSE
...and foo[["a"]]
is safer than foo$a
, since the latter uses partial matching and thus might also match a longer name:
x <- list(abc=4)
x$a # 4, since it partially matches abc
x[["a"]] # NULL, no match
[UPDATE] So, back to the question why exists('foo$a')
doesn't work. The exists
function only checks if a variable exists in an environment, not if parts of a object exist. The string "foo$a"
is interpreted literary: Is there a variable called "foo$a"? ...and the answer is FALSE
...
foo <- list(a=42, b=NULL) # variable "foo" with element "a"
"bar$a" <- 42 # A variable actually called "bar$a"...
ls() # will include "foo" and "bar$a"
exists("foo$a") # FALSE
exists("bar$a") # TRUE
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…