← SASnRRead SAS Datasets (.sas7bdat) in R

Introduction

Opening a .sas7bdat file is the very first thing a SAS programmer needs to learn when switching to R — if we cannot read our study data, nothing else matters.
SAS reads its own datasets natively through a LIBNAME and SET statement. R needs the haven package to do the same job.
haven::read_sas() preserves variable labels, formats, and data types, so the dataset in R looks the same as it did in SAS.
This lesson walks through the familiar SAS pattern first, then shows the equivalent in R so we can compare them side by side.
After running, we verify the column names, types, and labels all match what we expect from the original SAS dataset.

Read SAS Datasets (.sas7bdat) in R

SAS (Base SAS)+

*==============================================================================;
* Read SAS Datasets (.sas7bdat)
*==============================================================================;

*------------------------------------------------------------------------------;
* Step 1 - Point a libname to the folder containing the dataset
*------------------------------------------------------------------------------;
libname mylib "C:\clinical\data\sdtm";

*------------------------------------------------------------------------------;
* Step 2 - Read the dataset into WORK
*------------------------------------------------------------------------------;
data work.adsl;
    set mylib.adsl;
run;

*------------------------------------------------------------------------------;
* Step 3 - Confirm the structure
*------------------------------------------------------------------------------;
proc contents data=work.adsl;
run;

proc print data=work.adsl (obs=5);
run;

We start with the pattern every SAS programmer already knows — LIBNAME assigns a shortcut name to the folder where .sas7bdat files live.
The DATA/SET step copies the dataset from the permanent library into WORK so we can manipulate it freely.
If we only need to read, we can skip the copy and reference mylib.adsl directly in any procedure.
PROC CONTENTS shows the variable names, types, lengths, labels and formats — this is how we confirm the dataset loaded correctly.
PROC PRINT with (obs=5) gives us a quick look at the first five rows without flooding the log.
This SAS pattern is shown here purely as a reference — the real learning happens in the R tabs where we do the same thing with haven.

R (tidyverse)+

#==============================================================================
# Read SAS Datasets (.sas7bdat) in R
#==============================================================================

library(tidyverse)
library(haven)

#------------------------------------------------------------------------------
# Step 1 - Read the .sas7bdat file
#------------------------------------------------------------------------------
adsl <- read_sas("C:/clinical/data/sdtm/adsl.sas7bdat")

#------------------------------------------------------------------------------
# Step 2 - Quick look at the structure
#------------------------------------------------------------------------------
glimpse(adsl)

#------------------------------------------------------------------------------
# Step 3 - Check variable labels
#------------------------------------------------------------------------------
adsl %>%
  map_chr(~ attr(.x, "label") %||% "(no label)") %>%
  enframe(name = "variable", value = "label") %>%
  print(n = Inf)

#------------------------------------------------------------------------------
# Step 4 - Read only specific columns (saves memory)
#------------------------------------------------------------------------------
adsl_small <- read_sas(
  "C:/clinical/data/sdtm/adsl.sas7bdat",
  col_select = c(USUBJID, SITEID, SEX, AGE, TRT01P)
)

We load the haven package, which provides read_sas() for reading .sas7bdat files directly.
haven ships with the tidyverse ecosystem but it works perfectly fine in base R too — it has no dependency on dplyr or tibble beyond using them for output.
Note the forward slashes in the file path — R accepts forward slashes on Windows, and they are easier to read than double backslashes.
read_sas() preserves variable labels as attributes, keeps numeric and character types intact, and converts SAS dates to R date objects automatically.
glimpse() shows each column name, type and first few values in a compact layout — the tidyverse equivalent of PROC CONTENTS plus PROC PRINT.
To inspect the labels, we use map_chr() to pull the "label" attribute from every column and print them as a clean two-column table.
The col_select argument lets us read only the columns we need, which saves memory when datasets have hundreds of variables.
After running, we confirm that column names, types and labels match what PROC CONTENTS showed in SAS.

Happy learning!

Search lessons

Introduction

Read SAS Datasets (.sas7bdat) in R