*Copyright @ www.mycsg.in;

What does RETAIN do in a DATA step

• By default, SAS sets variables created in a DATA step to missing at the start of each new iteration unless they are read from a dataset or otherwise automatically retained
• The `retain` statement tells SAS to preserve the value of a variable from one iteration to the next
• This is especially useful for counters, running totals, and group-level calculations

Create input dataset

• The dataset `scores` contains test-level scores for each subject
• We will sort this dataset by subject and then use retained variables inside BY group processing examples

data scores; infile cards truncover; input usubjid $ item score; cards; 1001 1 93 1001 2 80 1001 3 80 1002 1 90 1002 2 95 1002 3 97 1003 1 100 1003 3 98 ; run; Copy Code View Log

• Each subject has one or more rows in the input dataset
• This structure is ideal for demonstrating retained counters and running statistics by subject

View Data

Find the total number of records for each student

• We sort by `usubjid` so BY group processing works correctly
• The retained variable `ntests` keeps its value across rows of the same subject
• When a new subject begins, `first.usubjid` resets the counter

proc sort data=scores; by usubjid; run; data output; set scores; by usubjid; retain ntests; if first.usubjid then ntests=1; else ntests=ntests+1; run; Copy Code View Log

• `ntests` shows a running count of rows within each subject
• Inspect the last row of each subject to identify the total number of records for that subject

View Data

Find the running total score for each subject

• The retained variable `total` carries forward the accumulated score value within each subject
• At the start of a new subject, the total is reset to the first score of that subject

proc sort data=scores; by usubjid; run; data output; set scores; by usubjid; retain total; if first.usubjid then total=score; else total=total+score; run; Copy Code View Log

• The variable `total` is a running cumulative score within each subject
• Check the last row for each subject to see the final total score

View Data

Find the highest score for each subject

• The retained variable `max_score` stores the highest value seen so far within the subject
• It is initialised on the first row of each subject, and updated only when a higher score is found

proc sort data=scores; by usubjid; run; data output; set scores; by usubjid; retain max_score; if first.usubjid then max_score=score; else if score gt max_score then max_score=score; run; Copy Code View Log

• `max_score` tracks the running maximum within each subject
• Look at the final row of each subject to confirm the highest score across that subject's tests

View Data

Find the lowest score for each subject

• The retained variable `min_score` works the same way, but keeps the smallest score seen so far
• This pattern is common for deriving minimum and maximum visit values within groups

proc sort data=scores; by usubjid; run; data output; set scores; by usubjid; retain min_score; if first.usubjid then min_score=score; else if score lt min_score then min_score=score; run; Copy Code View Log

• `min_score` tracks the running minimum within each subject
• Verify the final value on the last row of each subject against the subject's individual scores

View Data

Key points to remember

• `retain` preserves a variable's value across DATA step iterations
• `retain` is often used together with `first.` and `last.` BY group indicators
• Retained variables are useful for counters, running totals, running maximums, and running minimums