# First, you should be able to create a labelled spss vector with no labels# (That's a bit inelegant but it's expedient)
labelled_spss(c(1.2, 2.4), na_values=9)
labelled_spss(c(1.2, 2.4), double(), na_values=9)
# Also need to make sure the C++ code returns an object with that structure# Second you should be able to subset that object# (That's the root cause of the summary failure)x<- structure(c(1.2, 2.4), class= c("labelled_spss", "labelled"), na_values=9)
Would you mind providing an SPSS file that contains that variable so I can test?