Mixing Input Styles

Reading Free-Format Data

Mixing Input Styles

Evaluating your raw data and choosing the most appropriate input style is a very important task. You have probably worked with three primary input styles for reading raw data.

Input Style Reads

Column standard data values in fixed fields

Formatted nonstandard data values in fixed fields

List data values that are not arranged in fixed fields, but are separated by blanks or other delimiters

Input Style	Reads
Column	standard data values in fixed fields
Formatted	nonstandard data values in fixed fields
List	data values that are not arranged in fixed fields, but are separated by blanks or other delimiters

With some file layouts, you may need to mix input styles in the same INPUT statement in order to read the data correctly.
Look at the raw data file below and think about how to combine input styles to read these values.

1---+----10---+----20---+----30---+----40-

209-20-3721 07JAN78 41,983 SALES 2896

223-96-8933 03MAY86 27,356 EDUCATION 2344

232-18-3485 17AUG81 33,167 MARKETING 2674

251-25-9392 08SEP84 34,033 RESEARCH 2956

Column input is an appropriate choice for the first field because the values can be read as standard character values and are located in fixed columns.

The next two fields are also located in fixed columns, but the values require an informat. So, formatted input is a good choice here.

Values in the fourth field begin in column 28 but do not end in the same column. List input is appropriate here, but notice that some values are longer than eight characters. You need to use the : format modifier with an informat to read these values.

The last field does not have a common beginning or ending column, so list input is the best input style for these values.

Field Description

Starting Column

Field Width
Data Type
Input Style

Social Security #

1

11
character
column

Date of Hire

13

7
date
formatted

Annual Salary

21

6
numeric
formatted

Department
28

5 to 9
character list

Phone Extension

??

4
numeric
list

Field Description	Starting Column	Field Width	Data Type	Input Style
Social Security #	1	11	character	column
Date of Hire	13	7	date	formatted
Annual Salary	21	6	numeric	formatted
Department	28	5 to 9	character	list
Phone Extension	??	4	numeric	list

The INPUT statement to read the data should look like this:

data perm.mixed; infile rawdata; input SSN $ 1-11 @13 HireDate date7.
@21 Salary comma6. Department : $9. Phone; run; proc print data=perm.mixed; run;

When you submit the PRINT procedure, the output displays values for each variable.

Obs
SSN
HireDate Salary Department Phone

1 209-20-3721 6581 41983
SALES

2896

2 223-96-8933 9619 27356
EDUCATION

2344

3 232-18-3485 7899 33167 MARKETING
2674

4 251-25-9392 9017 34033
RESEARCH

2956

Obs	SSN	HireDate	Salary	Department	Phone
1	209-20-3721	6581	41983	SALES	2896
2	223-96-8933	9619	27356	EDUCATION	2344
3	232-18-3485	7899	33167	MARKETING	2674
4	251-25-9392	9017	34033	RESEARCH	2956