[R] average among one factor in a nested dataframe

[Junqian Gordon Xu]

I have two nested data frames:

a<-rnorm(6)
b<-rnorm(9)
f1<-c("x1","x2","x3"))
f2<-c("y1","y2")
id<-c(1:6)
a_df<-data.frame(cbind(id,f1,"y1",a))
id<-c(1:9)
b_df<-data.frame(cbind(id,f1,"y2",b))

I want to preserve id and f1, but want to collapse f2 and take the
corresponding mean values of a and b. Missing value in either dataframe
should be handled properly (i.e., just take the non-missing number
without dividing by 2).

I had a look at rowSum/Means and s/l/tapply, but couldn't figure out how
to handle this case cleanly. Any suggestions?

Thanks
Gordon