Replace values with missing values or fill missing values

Convert between missing values and explicit replacement codes:

• mvdecode() (na_if_value()): Replace specified codes with NA, like Stata's mvdecode x, mv(-98 -99).

• mvencode() (value_if_na()): Fill missing values with one replacement, like Stata's mvencode x, mv(0).

The aliases perform the same operations but name the replacement argument value instead of mv.

Usage

mvdecode(x, mv)

mvencode(x, mv)

na_if_value(x, value)

value_if_na(x, value)

Arguments

x

A vector.

mv, value

For mvdecode() and na_if_value(), the codes to replace with NA; any value equal to one of them becomes missing. For mvencode() and value_if_na(), a single replacement value for missing elements, cast to the type of x.

Value

A vector with the same type and size as x, with matching values replaced or missing values filled.

Details

Codes are matched with dplyr::na_if(); filling follows tidyr::replace_na(). Codes and replacements are cast to the type of x, so a lossy cast such as 1.5 for an integer vector is an error when applied. Filling a non-SPSS vector with no missing values returns it unchanged without casting the replacement. Both the input vector and the codes or replacement must be supplied. For SPSS labelled vectors (haven::labelled_spss()), user-defined missing codes and ranges also count as missing, matching their is.na() method; labels and missing-value metadata are retained.

See also

Other conversion functions: type_conversion

Examples

x <- c(1, -99, NA, 4, -98)
mvdecode(x, -99)
#> [1]   1  NA  NA   4 -98

mvdecode(x, c(-98, -99))
#> [1]  1 NA NA  4 NA

mvencode(x, 0)
#> [1]   1 -99   0   4 -98