Add `user_na` option to `zap_labels()`

[gorcha]

Currently zap_labels() converts user missing values in labelled_spss objects to NA.

This should be controllable by the user. I think adding a user_na option like read_sav() is the best approach, where FALSE (default) converts to NA and TRUE keeps them as valid values.

library(haven)
x <- labelled_spss(1:3, na_values = 1)
# Current behaviour
zap_labels(x)
#> [1] NA  2  3

# Desired behaviour
zap_labels(x, user_na = TRUE)
#> [1] 1  2  3

^{Created on 2021-09-15 by the reprex package (v2.0.1)}

[LuiNov]

Hey,

I have an issue reading a dta file with several missings. Numerics are -5:0 and labels are ("Missing:Other", "Not askey") etc. . How can I specify those to the function so I get the clean dataset ?????

I think it doesn't fit into this issue but I didn't know elsewhere to comment.... Thanks a lot :-)

[gorcha]

Hi @LuiNov - I've just moved this to its own issue at #645, shall reply over there.