Replaces tagged NAs with their original missing value codes. Works with
data imported via read_spss() (with tag_na = TRUE) or any reader
that used the tag_na parameter (read_stata(), read_sas(),
read_xpt()). For native Stata/SAS tagged NAs (e.g., .a, .A) that
have no numeric codes to recover, use strip_tags() instead.
Value
The input with tagged NAs that have numeric codes replaced by
their original values (e.g., -9, -8, -42). System NAs (untagged)
remain NA. Value labels and the variable label are kept (labels of
missing types are attached to their codes again), so the result is
haven_labelled when the input was. For native Stata/SAS tagged NAs
(no numeric codes), falls back to strip_tags() behavior with a
warning. A data frame is returned with the selected columns converted.
See also
read_spss(), read_stata(), read_sas(), na_frequencies(),
strip_tags()
Other data-import:
na_frequencies(),
read_por(),
read_sas(),
read_spss(),
read_stata(),
read_xlsx(),
read_xpt(),
strip_tags()
Examples
# \donttest{
if (requireNamespace("haven", quietly = TRUE)) {
# Tag -9/-8 as missing, then recover the original codes
x <- set_na(c(1, 2, -9, 3, -8), -9, -8)
untag_na(x) # -9 and -8 are back
}
#> <labelled<double>[5]>
#> [1] 1 2 -9 3 -8
# }