How to test if list element exists?
Solution 1:
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
Solution 2:
The best way to check for named elements is to use exist()
, however the above answers are not using the function properly. You need to use the where
argument to check for the variable within the list.
foo <- list(a=42, b=NULL)
exists('a', where=foo) #TRUE
exists('b', where=foo) #TRUE
exists('c', where=foo) #FALSE
Solution 3:
One solution that hasn't come up yet is using length, which successfully handles NULL. As far as I can tell, all values except NULL have a length greater than 0.
x <- list(4, -1, NULL, NA, Inf, -Inf, NaN, T, x = 0, y = "", z = c(1,2,3))
lapply(x, function(el) print(length(el)))
[1] 1
[1] 1
[1] 0
[1] 1
[1] 1
[1] 1
[1] 1
[1] 1
[1] 1
[1] 1
[1] 3
Thus we could make a simple function that works with both named and numbered indices:
element.exists <- function(var, element)
{
tryCatch({
if(length(var[[element]]) > -1)
return(T)
}, error = function(e) {
return(F)
})
}
If the element doesn't exist, it causes an out-of-bounds condition caught by the tryCatch block.
Solution 4:
Here is a performance comparison of the proposed methods in other answers.
> foo <- sapply(letters, function(x){runif(5)}, simplify = FALSE)
> microbenchmark::microbenchmark('k' %in% names(foo),
is.null(foo[['k']]),
exists('k', where = foo))
Unit: nanoseconds
expr min lq mean median uq max neval cld
"k" %in% names(foo) 467 933 1064.31 934 934 10730 100 a
is.null(foo[["k"]]) 0 0 168.50 1 467 3266 100 a
exists("k", where = foo) 6532 6998 7940.78 7232 7465 56917 100 b
If you are planing to use the list as a fast dictionary accessed many times, then the is.null
approach might be the only viable option. I assume it is O(1), while the %in%
approach is O(n)?
Solution 5:
A slight modified version of @salient.salamander , if one wants to check on full path, this can be used.
Element_Exists_Check = function( full_index_path ){
tryCatch({
len_element = length(full_index_path)
exists_indicator = ifelse(len_element > 0, T, F)
return(exists_indicator)
}, error = function(e) {
return(F)
})
}