dplyr inner_join与字符列上的NAs

Question

I have two equal data frames

a <- c(1,2,3)
b <- c(3,2,1)
c <- c('a','b',NA)

df1 <- data.frame(a=a, b=b, c=c, stringsAsFactors=FALSE)
df2 <- data.frame(a=a, b=b, c=c, stringsAsFactors=FALSE)

I would like to use dplyr::inner_join to

"return all rows from x where there are matching values in y, and all columns from x and y" dplyr documentation

(which is everything as they are equal) but it doesn't seem to work with an NA in column c (type chr ). Is this standard behaviour to not join on the NA s?

For example

library(dplyr)
> inner_join(df1, df2)
Joining by: c("a", "b", "c")
  a b c
1 1 3 a
2 2 2 b

doesn't join on the NA . However, I would like it to return the same as merge

> merge(df1, df2)
  a b    c
1 1 3    a
2 2 2    b
3 3 1 <NA>

Have I misunderstood how inner_join works in this instance and is this behaving as described?

Further Detail

inner_join matches NA on a numeric column

a <- c(1,2,3)
b <- c(3,2,NA)
c <- c('a','b','c')

df1 <- data.frame(a=a, b=b, c=c, stringsAsFactors=FALSE)
df2 <- data.frame(a=a, b=b, c=c, stringsAsFactors=FALSE)

> inner_join(df1, df2)
Joining by: c("a", "b", "c")
  a  b c
1 1  3 a
2 2  2 b
3 3 NA c

Edit

As @thelatemail points out, inner_join also works as merge when the NA is in a factor column

df1 <- data.frame(a=a, b=b, c=c, stringsAsFactors=T)
df2 <- data.frame(a=a, b=b, c=c, stringsAsFactors=T)
inner_join(df1, df2)
Joining by: c("a", "b", "c")
  a b    c
1 1 3    a
2 2 2    b
3 3 3 <NA>

Edit 2

Thanks to @shadow for pointing out this is a known issue here and here

Answer 1

This issue was occurring in version 0.4.1. This is now fixed in version 0.4.2:

sessionInfo()
...
other attached packages:
[1] dplyr_0.4.2
...

> inner_join(df1, df2)
Joining by: c("a", "b", "c")
  a b    c
1 1 3    a
2 2 2    b
3 3 1 <NA>

Check with merge:

> merge(df1, df2)
  a b    c
1 1 3    a
2 2 2    b
3 3 1 <NA>

> all.equal(inner_join(df1, df2), merge(df1, df2))
Joining by: c("a", "b", "c")
[1] TRUE