Merge with NA in factor key copies NA row

[stewbasic]

I'm using data.table v1.9.4 on R 3.0.2.

I'm merging two data.tables with a key of type factor containing an NA entry. For keys appearing only in the first table, the columns from the second table don't contain NA, as expected from ?merge, but instead get copied from the NA row. See row 3, column v.y in the following example:

x<-data.table(k=as.factor(c(NA,1,2)),v=c(0,1,2))
y<-data.table(k=as.factor(c(NA,1,3)),v=c(0,1,3))
merge(x,y,all=T,by="k")
#     k v.x v.y
#1: NA   0   0
#2:  1   1   1
#3:  2   2   0
#4:  3  NA   3

This is very similar to the issue reported in this SO post, but it should be fixed as of v1.8.11 according to the comment there.

[arunsrinivasan]

@stewbasic thanks for the nice minimal example.