CODEWITHSUNDEEP
r
Daniel V
Daniel V

Reputation: 1386

Different vector lengths when using NA

I'm getting interesting results from the following and am not sure why.

match(3, 1:2)

returns NA

c(1, 3)[NA]

returns NA NA

c(1, 3)[match(3, 1:2)]

returns NA

What's going on here? Why are the second and third results different lengths? Intuitively, substituting match(3, 1:2) with its result NA should come up with the same result but it doesn't. What triggers that?

Upvotes: 2

Views: 39

Answers (1)

akrun
akrun

Reputation: 887148

There is recycling effect. In the case of match, if there is no matches, by default nomatch = NA_integer

c(1, 3)[match(3, 1:2)]
#[1] NA

while the OP used a NA_logical_ while not specifying the type and it just recycles the NA to the length of the vector i.e. 2. If we do the type conversion, will get the same output

c(1, 3)[as.logical(match(3, 1:2))]
#[1] NA NA

Same thing can be replicated in reverse

c(1, 3)[NA_integer_]
#[1] NA

The reason is mentioned in ?NA

NA is a logical constant of length 1 which contains a missing value indicator. NA can be coerced to any other vector type except raw. There are also constants NA_integer_, NA_real_, NA_complex_ and NA_character_ of the other atomic vector types which support missing values: all of these are reserved words in the R language.

Logical computations treat NA as a missing TRUE/FALSE value, and so may return TRUE or FALSE if the expression does not depend on the NA operand.

Upvotes: 2

Related Questions