← SASnRSort Order with Missing Values

Introduction

Sorting becomes trickier the moment missing values enter the picture — do NAs belong at the top, at the bottom, or should they be dropped?
SAS PROC SORT places missing values first in ascending order and last in descending order.
Tidyverse arrange() always pushes NAs to the bottom by default, regardless of sort direction, so we steer them explicitly.
Base R order() drops NAs by default; we can change that with the na.last argument or a small trick on missingness.
This lesson sorts a small dataset where two rows have a missing score, so the placement of the NAs is easy to see in each language.

Sort Order with Missing Values

SAS (Base SAS)+

data sort_missing;
    input id $ score;
datalines;
A 85
B .
C 90
D 70
E .
F 95
G 80
;

run;

proc sort data=sort_missing out=sorted;
    by score;
run;

In SAS, missing values are treated as smaller than any number in ascending order.
proc sort data=sort_missing out=sorted; by score; sorts on score ascending and writes the sorted rows to SORTED.
Because missing is the smallest value, the two rows with . come out first, followed by the numeric scores in increasing order.
For descending sort, by descending score; flips the order and the missing-value rows drop to the bottom.
This behaviour is built into PROC SORT — we never have to code a special rule to get NAs first.
After running, open SORTED and confirm the two missing rows sit at the top.

R (tidyverse)+

library(tidyverse)


data <- tibble( 
id = c("A", "B", "C", "D", "E", "F", "G"), 
score = c(85, NA, 90, 70, NA, 95, 80) 
)


data_sorted <- data %>%
   arrange(!is.na(score), score)

By default, arrange() from dplyr pushes NAs to the bottom regardless of sort direction — the opposite of the SAS ascending default.
arrange(!is.na(score), score) is a two-key trick: sort first on a logical flag, then on score itself.
!is.na(score) is FALSE for NA rows and TRUE for non-NA rows — FALSE sorts before TRUE, so NAs land on top.
For the opposite placement, use arrange(is.na(score), desc(score)) to keep NAs at the bottom while sorting values descending.
Both forms are explicit, which makes the intent obvious to the next person reading the code.
We run and verify the two NA rows are where we expect them to be.

Happy learning!

Search lessons

Introduction

Sort Order with Missing Values