← SASnRCreate a macro variable and use it in open code

Introduction

Programs often need a reusable scalar — a treatment label to filter on, a study cut-off date, a current version string — that may change between runs.
The pattern is: define the value once at the top and reference it wherever it is needed.
SAS has a dedicated macro namespace: %let defines a macro variable and &name. references it.
R has no macro layer — a plain variable in the environment serves the same purpose, referenced by its bare name.
In dplyr, column names and environment variables are disambiguated automatically, so referencing the filter value is as simple as writing its name.
This lesson filters an ADSL sample down to subjects on the Placebo arm, showing the same parameterised pattern in SAS, tidyverse, and base R.

Create a macro variable and use it in open code

SAS (Base SAS)+

*=============================================================================;
* Create sample ADSL data;
*=============================================================================;

data adsl;
    input usubjid $ trtp $ age sex $;
    datalines;
SUBJ001 Placebo 67 M
SUBJ002 DrugA   59 F
SUBJ003 Placebo 72 F
SUBJ004 DrugB   65 M
SUBJ005 Placebo 60 F
;
run;

*=============================================================================;
* Define macro variable and use in filter;
*=============================================================================;

%let trtp_filter = Placebo;

data adsl_filtered;
    set adsl;
    if trtp = "&trtp_filter.";
run;

To make a reusable scalar in SAS, we define a macro variable with %let and reference it with the &name. syntax.
%let trtp_filter = Placebo; creates a macro variable whose value is the text "Placebo" — no quotes needed in the assignment.
In the DATA step, if trtp = "&trtp_filter."; resolves to if trtp = "Placebo"; before SAS compiles the step.
We put the reference inside double quotes so SAS still sees a quoted string literal after macro resolution. Single quotes would prevent resolution — the &trtp_filter. would be taken as literal text.
The trailing dot terminates the macro variable name — helpful when the next character could be mistaken for part of the name.
This is the go-to pattern for parameterising a program by study, treatment, cut-off date, or any other single value that might change between runs.
After running, we check adsl_filtered and confirm it contains only the Placebo rows from adsl.

R (tidyverse)+

library(tidyverse)

#==============================================================================
# Create sample ADSL data
#==============================================================================


adsl <- tribble(
  ~usubjid, ~trt01p,     ~age, ~sex,
  "SUBJ001", "Placebo", 67,   "M",
  "SUBJ002", "DrugA",    59,   "F",
  "SUBJ003", "Placebo", 72,   "F",
  "SUBJ004", "DrugB",    65,   "M",
  "SUBJ005", "Placebo", 60,   "F"
)

#==============================================================================;
#Create an object and use it's value in filter function;
#==============================================================================;

trt_filter <- "Placebo"

adsl_filtered <- adsl %>% 
  filter(trt01p==trt_filter)

R does not have a separate macro namespace — regular variables do the same job.
We assign trt_filter <- "Placebo" with a normal assignment operator. The value lives in the R environment alongside our data.
Inside filter(), we reference the variable by name with no special syntax: filter(trt01p == trt_filter).
dplyr quietly distinguishes between column names (bare identifiers that exist in the tibble) and environment objects (named values like trt_filter) — no & or resolution step needed.
This is cleaner than the SAS macro approach for the simple single-value case: one assignment, one reference, and it plays nicely with autocomplete and tooling.
After running, we confirm adsl_filtered only contains the Placebo rows.

Happy learning!

Search lessons

Introduction

Create a macro variable and use it in open code