Return FALSE for duplicated NA values when using the function duplicated()

Question

just wondering why duplicated behaves the way it does with NAs:

> duplicated(c(NA,NA,NA,1,2,2))
[1] FALSE  TRUE  TRUE FALSE FALSE  TRUE

where in fact

> NA == NA
[1] NA

is there a way to achieve that duplicated marks NAs as false, like this?

> duplicated(c(NA,NA,NA,1,2,2))
[1] FALSE  FALSE  FALSE FALSE FALSE  TRUE

`duplicated` marks the second (and third, and fourth, etc.) occurrences as duplicated, but not the first. You can use `is.na()` to do what you ask. — Andrie, Nov 27 '12 at 11:46
thanks. The main question is why it makes sense to mark NAs as duplicates. — jamborta, Nov 27 '12 at 11:51

score 24 · Accepted Answer · answered Nov 27 '12 at 11:52

You use the argument incomparables for the function duplicated like this :

> duplicated(c(NA,NA,NA,1,2,2))
[1] FALSE  TRUE  TRUE FALSE FALSE  TRUE
> duplicated(c(NA,NA,NA,1,2,2),incomparables=NA)
[1] FALSE FALSE FALSE FALSE FALSE  TRUE

It determines the values that cannot be compared (in this case NA) and returns FALSE for those values. See also ?duplicated

Return FALSE for duplicated NA values when using the function duplicated()

1 Answers1

Related