CODEWITHSUNDEEP
rsubset
lost
lost

Reputation: 1669

In R, negative subsetting produces unexpected results with slightly larger numbers

In R,

print((1:20)[-(15:18)])

prints the list of integers from 1 to 20 excepting 15 through 18, as expected.

But

print((1983:2010)[-(1992:1994)])

prints the list of integers from 1983 to 2010, including 1992 through 1994. The negative subsetting is not working.

What's going on here? I know R can have trouble with extremely large numbers, but I would not consider these numbers extremely large.

Upvotes: 1

Views: 1296

Answers (2)

akrun
akrun

Reputation: 887971

This is based on the numeric index. In the second example, we don't have a vector of length > 1994, but in the former case, there is

length(1983:2010)
#[1] 28

The way to deal with this to use %in%

v1 <- 1983:2010
v1[!v1 %in% 1992:1994]

Or use the index based on the original length of the vector

v1[-((length(v1)-2):length(v1))]

Upvotes: 1

neilfws
neilfws

Reputation: 33822

The values in square brackets refer to the position of the number in the vector, not the number itself.

In your first example, numbers 15-18 also occupy the positions 15-18.

In your second example, numbers 1992-1994 don't occupy positions 1992-1994. They occupy positions 10-12. So you could use:

print((1983:2010)[-(10:12)])

Upvotes: 0

Related Questions