← SASnRReplicate PROC FORMAT Using Named Vectors in R

Introduction

Clinical data often travels as numeric codes — sex = 1/2, race = 1/2/3/4 — but reports and SDTM want the human-readable labels.
The mapping from code to label is reusable: define it once and apply it wherever the coded column appears.
SAS defines the mapping with proc format and applies it with put(code, fmt.) to derive a labelled column.
Tidyverse and base R both use the same trick: a named character vector, indexed by the string form of the code.
Named-vector lookups are fast, easy to read, and need no extra package — the closest R idiom to SAS formats.
This lesson maps Sex codes 1 and 2 to "Male" and "Female" and adds the label as a new column sexc.

Replicate PROC FORMAT Using Named Vectors in R

SAS (Base SAS)+

* Create format using PROC FORMAT;
proc format;
    value sexfmt
    1 = "Male"
    2 = "Female"
    ;
run;

data dm;
  input usubjid $ sex;
  datalines;
01 1
02 2
03 1
04 2
;
run;

* Apply the format and create labeled variable;
data dm_labeled;
    set dm;
    sexc = put(sex, sexfmt.);
run;

In SAS, the code-to-label mapping is declared with proc format and then applied with put().
proc format; value sexfmt 1 = "Male" 2 = "Female"; run; defines a reusable format called sexfmt.
The format is stored in the current format catalogue and is available to any DATA step or PROC until the session ends.
sexc = put(sex, sexfmt.); translates the numeric code into its label — the trailing dot on sexfmt. tells SAS it is a user-defined format.
Unlike a merge-lookup, the format costs almost nothing to evaluate row by row and stays attached to the column as metadata.
We run and open DM_LABELED; the new sexc column should read "Male" for sex = 1 and "Female" for sex = 2.

R (tidyverse)+

library(tidyverse)

# Sample tibble with coded values

dm <- tibble( usubjid = c("01", "02", "03", "04"), sex = c(1, 2, 1, 2) )

# Define named vector format

sex_fmt <- c("1" = "Male", "2" = "Female")

# Apply the format using mutate

dm_labeled <- dm %>%

mutate( sexc = sex_fmt[as.character(sex)] )

In R, the closest thing to a SAS format is a named character vector: the names are the codes and the values are the labels.
sex_fmt <- c("1" = "Male", "2" = "Female") defines the mapping; both names and values are strings.
mutate(sexc = sex_fmt[as.character(sex)]) looks up each code in the named vector and pulls out the matching label.
as.character(sex) is important because the names of the vector are strings — indexing a named vector with numeric values would be by position, not by name.
Codes not present in the mapping return NA, so unexpected values are flagged rather than silently mislabelled.
Named vectors cover most use cases; dplyr::recode() or dplyr::case_when() are alternatives when the mapping logic is more complex than a flat lookup.

Happy learning!

Search lessons

Introduction

Replicate PROC FORMAT Using Named Vectors in R