Normalize Missing Value Representations in SRTR Data — srtr_normalize

Replaces common non-standard representations of missing values (e.g., -1, 999, "Unknown") with either NA or a user-specified string (e.g., "Missing"). This includes both explicit missing codes (provided via missing_vals) and, if replacement is not NA, any existing NA values in the specified columns.

Usage

srtr_normalize_missing(
  df,
  missing_vals = NULL,
  replacement = NA,
  verbose = FALSE
)

Arguments

df: A data frame from an SRTR SAF file.
missing_vals: An optional named list. The srtr_normalize_missing() function defaults to an internal list of variables to normalize and values to convert to the specified replacement value. Each name corresponds to a column in df, and each element is a vector of values to treat as missing in that column. Any values matched will be replaced by replacement. Columns not present in df are silently skipped.
replacement: The value to use in place of the missing representations. Defaults to NA. If a string is provided (e.g., "Missing"), both missing codes and NA values will be replaced. If the column is a factor and replacement is not NA, the replacement will be added as a new factor level.
verbose: Logical. If TRUE, the function prints a message for each column it modifies, including skipped columns that are not found in df.

Value

A modified version of df, with all specified missing values (and possibly NAs) replaced by replacement.

Details

This function is helpful for harmonizing SRTR datasets, which may use multiple codes to represent missingness (e.g., -1 for numeric fields, "U" or "Unknown" for character or factor fields).

Examples

if (FALSE) {
df <- load_srtr_file("TX_LI")

# Use default SRTR-wide missing definitions
srtr_normalize_missing(df)

# Explicitly define missing values and convert all missing to "Missing"
missing_vals <- list(
  REC_HIV_STAT = c("U", ""),
  REC_HCV_STAT = c("U", ""),
)
srtr_normalize_missing(df, missing_vals, replacement = "Missing")
}