healthcareai (version 2.3.0)

make_na: Replace missingness values with NA and correct columns types

Description

This function replaces given missingness values with NA in a given dataframe or tibble. Numeric vectors that were originally loaded as character or factor vectors (because of missingness values in the column), are also converted to numeric vectors when values are replaced.

Usage

make_na(d, to_replace, drop_levels = TRUE)

Arguments

d

A dataframe or tibble

to_replace

A value or vector of values that will be replaced with NA

drop_levels

If TRUE (default) unused factor levels are dropped

Value

A tibble where the missing value/values is/are replaced with NA, columns that only have numbers left are coerced to numeric type

Examples

# NOT RUN {
dat <- data.frame(gender = c("male", "male", "female", "male", "missing"),
              name = c("Paul", "Jim", "Sarah", "missing", "Alex"),
              weight = c(139, 0, 193, 158, 273))

# Replace "missing" in `dat`
make_na(dat, "missing")

# If there are multiple missing values, pass them through a vector.
dat <- data.frame(gender = c("male", "??", "female", "male", "NULL"),
              age = c(64, 52, 75, "NULL", 70),
              weight = c(139, 0, 193, "??", 273),
              stringsAsFactors = FALSE)

make_na(dat, c("??", "NULL"))

# Run `missingness()` to find possible missingness values in `dat`. It will
# suggest the default implementation of `make_na` to replace all found
# missingness values (the suggested default implementation for this example
# is `make_na(dat, c("??", "NULL"))`).
missingness(dat)
make_na(dat, c("??", "NULL"))

# Note: In this last example, `age` should be loaded as a numeric vector, but
# since "NULL" is present, it is stored as a character vector. When "NULL" is
# replaced, `age` will be converted to a numeric vector.

# }