1.4. Migrating from SAS to R: A Skill Conversion Guide

1.4.7 Slicing Data in SAS vs. R

Slicing refers to selecting specific rows from a dataset—by position, condition, rank, group, or time.

• In SAS, this is usually done using _N_, FIRSTOBS=, OBS=, BY-group logic, or PROC SQL/SORT.
• In R, slicing is handled by functions like slice(), filter(), slice_head(), slice_max(), and so on—especially with dplyr.

Each use case below is illustrated with:

• Concept
• SAS code + explanation
• R code + explanation
• Input table
• Expected output

1. Select Specific Row Numbers

Concept Pick rows 2, 5, and 7.

SAS Code

data sales_data;
  input Region $ Sales Profit;
  datalines;
East 1200 200
West 1300 210
East 1400 220
North 1350 215
South 1250 205
West 1280 202
East 1340 208
;
run;

data want;
  set sales_data;
  if _n_ in (2,5,7);
run;

SAS Code Explanation

• data want;
  • Starts a new data step to create the dataset want.
• set sales_data;
  • Reads each row from sales_data sequentially.
• if _n_ in (2,5,7);
  • _n_ is an automatic variable indicating the current row number.
  • This condition keeps only rows 2, 5, and 7.
• run;
  • Ends the data step and writes the output.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  slice(c(2, 5, 7))

R Code Explanation

• sales_data %>%
  • Passes the sales_data data frame into the next function using the pipe.
• slice(c(2, 5, 7))
  • Selects rows at positions 2, 5, and 7.
  • Returns a new data frame with only those rows.

Input

Output

2. Slice a Range of Rows

Concept Select rows 3 through 6.

SAS Code

proc print data=sales_data(firstobs=3 obs=6);
run;

SAS Code Explanation

• proc print data=sales_data(firstobs=3 obs=6);
  • firstobs=3 tells SAS to start at row 3.
  • obs=6 tells SAS to stop at row 6.
  • Only rows 3, 4, 5, and 6 are printed.
• run;
  • Executes the procedure.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  slice(3:6)

R Code Explanation

• sales_data %>%
  • Pipes the data frame to the next function.
• slice(3:6)
  • Selects rows 3, 4, 5, and 6 by position.

Output

3. Filter Then Slice

Concept Filter on a condition, then pick top rows from the filtered set.

SAS Code

data high_profit;
  set sales_data;
  if Profit > 205 and _n_ <= 2;
run;

SAS Code Explanation

• data high_profit;
  • Creates a new dataset for high-profit rows.
• set sales_data;
  • Reads each row from sales_data.
• if Profit > 205 and _n_ <= 2;
  • Keeps only rows where Profit is greater than 205 and the row number is 1 or 2 (after filtering).
• run;
  • Ends the data step.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  filter(Profit > 205) %>%
  slice(1:2)

R Code Explanation

• filter(Profit > 205)
  • Filters rows where Profit is greater than 205.
• slice(1:2)
  • After filtering, selects the first two rows from the result.

Output

4. Top N Rows by Value

Concept Pick top 3 rows by Sales.

SAS Code

proc sort data=sales_data out=sorted;
  by descending Sales;
run;

data top3;
  set sorted(obs=3);
run;

SAS Code Explanation

• proc sort data=sales_data out=sorted; by descending Sales; run;
  • Sorts sales_data in descending order by Sales and saves as sorted.
• data top3; set sorted(obs=3); run;
  • Reads only the first 3 rows from the sorted data set into top3.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  arrange(desc(Sales)) %>%
  slice_head(n = 3)

R Code Explanation

• arrange(desc(Sales))
  • Sorts the data frame in descending order by Sales.
• slice_head(n = 3)
  • Selects the first 3 rows from the sorted data.

Output

5. First Row per Group

Concept Pick first observation within each Region.

SAS Code

proc sort data=sales_data; by Region; run;

data first_in_group;
  set sales_data;
  by Region;
  if first.Region;
run;

SAS Code Explanation

• proc sort data=sales_data; by Region; run;
  • Sorts the data by Region so that group boundaries are contiguous.
• data first_in_group; set sales_data; by Region; if first.Region; run;
  • For each group of Region, keeps only the first row (where first.Region is true).

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  group_by(Region) %>%
  slice(1)

R Code Explanation

• group_by(Region)
  • Groups the data by Region.
• slice(1)
  • Selects the first row from each group.

Output

6. Drop Rows by Position

Concept Drop rows 1 and 7.

SAS Code

data want;
  set sales_data;
  if _n_ not in (1,7);
run;

SAS Code Explanation

• data want; set sales_data; if _n_ not in (1,7); run;
  • Reads all rows except those at positions 1 and 7.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  slice(-c(1, 7))

R Code Explanation

• slice(-c(1, 7))
  • Negative indices in slice() drop the specified rows.

7. Row with Maximum Value

Concept Get the row with the highest Profit.

SAS Code

proc sql outobs=1;
  select * from sales_data
  order by Profit desc;
quit;

SAS Code Explanation

• proc sql outobs=1;
  • Limits the output to 1 row.
• select * from sales_data order by Profit desc;
  • Orders by Profit descending, so the first row is the maximum.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  slice_max(Profit, n = 1)

R Code Explanation

• slice_max(Profit, n = 1)
  • Selects the row with the maximum value of Profit.

8. Top-N Rows per Group

Concept Get top 2 Sales rows within each Region.

SAS Code

proc sort data=sales_data;
  by Region descending Sales;
run;

data top2;
  set sales_data;
  by Region;
  retain counter 0;
  if first.Region then counter = 1;
  else counter + 1;
  if counter <= 2;
run;

SAS Code Explanation

• proc sort data=sales_data; by Region descending Sales; run;
  • Sorts by Region and then by Sales descending within each region.
• retain counter 0;
  • Initializes a counter variable.
• if first.Region then counter = 1; else counter + 1;
  • Resets counter at the start of each group, increments otherwise.
• if counter <= 2;
  • Keeps only the top 2 rows per group.

R Code

library(dplyr)
sales_data <- data.frame(
  Region = c("East", "West", "East", "North", "South", "West", "East"),
  Sales = c(1200, 1300, 1400, 1350, 1250, 1280, 1340),
  Profit = c(200, 210, 220, 215, 205, 202, 208)
)
sales_data %>%
  group_by(Region) %>%
  slice_max(Sales, n = 2)

R Code Explanation

• group_by(Region)
  • Groups by Region.
• slice_max(Sales, n = 2)
  • Selects the top 2 rows by Sales within each group.

9. Bottom-N per Group

Concept Pick row with lowest Profit in each Category.

SAS Code

data sales_data;
  input Category $ Profit;
  datalines;
A 100
A 90
B 80
B 70
;
run;

proc sort data=sales_data;
  by Category Profit;
run;

data bottom1;
  set sales_data;
  by Category;
  if first.Category;
run;

SAS Code Explanation

• data sales_data;
  • Creates a dataset with Category and Profit.
• proc sort data=sales_data; by Category Profit; run;
  • Sorts by Category and then by Profit ascending.
• if first.Category;
  • Keeps the first row in each category, which is the lowest Profit.

R Code

library(dplyr)
sales_data <- data.frame(
  Category = c("A", "A", "B", "B"),
  Profit = c(100, 90, 80, 70)
)
sales_data %>%
  group_by(Category) %>%
  slice_min(Profit, n = 1)

R Code Explanation

• group_by(Category)
  • Groups by Category.
• slice_min(Profit, n = 1)
  • Selects the row with the minimum Profit in each group.

10. Time-Based Slicing (Last 30 Days)

Concept Keep rows where the Date is within the last 30 days.

SAS Code

data sales_data;
  input Date : yymmdd10. Sales;
  format Date yymmdd10.;
  datalines;
2023-05-01 1000
2023-06-15 1100
2023-07-01 1200
2023-07-28 1500
;
run;

data recent;
  set sales_data;
  if Date >= today() - 30;
run;

SAS Code Explanation

• data sales_data;
  • Creates a dataset with Date and Sales.
• input Date : yymmdd10. Sales;
  • Reads Date in YYYY-MM-DD format.
• format Date yymmdd10.;
  • Sets the display format for Date.
• data recent; set sales_data; if Date >= today() - 30; run;
  • Keeps only rows where Date is within the last 30 days from today.

R Code

library(dplyr)
sales_data <- data.frame(
  Date = as.Date(c("2023-05-01", "2023-06-15", "2023-07-01", "2023-07-28")),
  Sales = c(1000, 1100, 1200, 1500)
)
sales_data %>%
  filter(Date >= Sys.Date() - 30)

R Code Explanation

• filter(Date >= Sys.Date() - 30)
  • Keeps rows where Date is within the last 30 days.
  • Make sure Date is of type Date in R.

Input

Output (assuming today = 2023-07-30)

Summary: SAS vs. R for Slicing

**Resource download links**

1.4.7.-Slicing-Data-in-SAS-vs-R.zip