Transforming Data with SAS Functions
Modifying Character Values with Functions

TRIM Function The TRIM function enables you to remove trailing blanks from character values. To learn about the TRIM function, let's complete a modification to the data set Hrd.Temp that requires this function. The data set Hrd.Temp contains four address variables, Address, City, State, and Zip.

SAS Data Set Hrd.Temp

Agency	ID	Name	Address	City	State	Zip	Phone	BeginDate	EndDate	PayRate	Days	Hours
Administrative Support, Inc.	F274	CICHOCK, ELIZABETH MARIE	65 ELM DR	CARY	NC	27513	6224549	14567	14621	10	11	88
Administrative Support, Inc.	F101	BENINCASA, HANNAH LEE	11 SUN DR	CARY	NC	27513	6223251	14524	14565	8	25	200

You need to create one address variable that contains the values of the three variables Address, City, and Zip. (Because all temporary employees are hired locally, the value of State does not need to be included in the new variable.)

SAS Data Set Hrd.Newtemp

Agency	ID	Name	NewAddress	Phone	BeginDate	EndDate	PayRate	Days	Hours
Administrative Support, Inc.	F274	CICHOCK, ELIZABETH MARIE	65 ELM DR, CARY, 27513	6224549	14567	14621	10	11	88
Administrative Support, Inc.	F101	BENINCASA, HANNAH LEE	11 SUN DR, CARY, 27513	6223251	14524	14565	8	25	200

Writing the DATA step to create this new variable is easy. You include an assignment statement that contains the concatenation operator, ||, as shown below. data hrd.newtemp(drop=address city state zip); set hrd.temp; NewAddress=address||', '||city||', '||zip; run; The concatenation operator (| |) enables you to concatenate character values. In this assignment statement, the character values of Address, City, and Zip are concatenated with two character constants consisting of a comma and a blank. The commas and blanks are needed to separate the street, city, and ZIP code values. The length of NewAddress is the sum of the length of each variable and constant that is used to create the new variable. Notice that this DATA step drops the original address variables from the new data set. When the DATA step is executed, you notice that the values of NewAddress do not appear as expected. The values of the new variable contain embedded blanks.

SAS Data Set Hrd.Newtemp

NewAddress
65 ELM DRIVE                    , CARY           , 27513
11 SUN DRIVE                    , CARY           , 27513
712 HARDWICK STREET             , CHAPEL HILL    , 27514
5372 WHITEBUD ROAD              , RALEIGH        , 27612

These blanks appear in the values of NewAddress because the values of the original address variables contained trailing blanks. Whenever the value of a character variable does not match the length of the variable, SAS software pads the value with trailing blanks.

Address length=32	City length=15	Zip length=5
65 ELM DRIVE····················	RALEIGH········	27612
11 SUN DRIVE····················	DURHAM·········	27612
712 HARTWICK STREET·············	CHAPEL HILL····	27514

Notice that the variable Zip does not contain trailing blanks. So, when the original address values are concatenated to create NewAddress, the trailing blanks in the original values are included in the values of the new variable.

NewAddress length=60
65 ELM DRIVE····················, RALEIGH········, 27612
11 SUN DRIVE····················, DURHAM·········, 27612
712 HARTWICK STREET·············, CHAPEL HILL····, 27514