How do I change a variable length (when using an infile statement)?

[vagnozzia408]

Hi, there! I am still becoming familiar with programming in SAS Studio 3.1, and am currently in the process of getting SAS to read CSV files that were created from Excel data spreadsheets. The file I'm using ("MyData_AMV.csv") is being imported and read as a table (work.MyData), which is great! But the values in the columns containing character string variables are being truncated to 8 characters, and I want to specify certain columns as having more characters in a string. For example, I want the "LastName" column to have a length of up to 20 characters.

data work.MyData;
	infile 'C:\\Users\Public\Documents\MyData_AMV.csv'
		dlm=',' firstobs=2;
	input StudentID LastName $ FirstName $ Age Gender $ Race $ InstitutionType $ InstitutionName $ 
                HighSchoolGPA Term $ CourseSection $ LetterGrade $ CurrentStatus $;
run;

Any suggestions on how to fix this problem would be much appreciated!

Thank you!

~AMV

[ballardw · Accepted solution]

SAS defaults to 8 characters for unspecified input.

Use an INFORMAT statement BEFORE the variable is input to assign length is easiest.

data work.MyData;
	infile 'C:\\Users\Public\Documents\MyData_AMV.csv'
		dlm=',' firstobs=2;
   informat LastName $20.;
   Informat FirstName $20.;
	input StudentID LastName $ FirstName $ Age Gender $ Race $ InstitutionType $ InstitutionName $ 
                HighSchoolGPA Term $ CourseSection $ LetterGrade $ CurrentStatus $;
run;

One hint for the future, you can use proc import with CSV and provide a largish GUESSINGROWS option value. Proc Import actually creates Datastep code that you can copy fromt the log and edit for anything you want to change.

[vagnozzia408]

That fixed the problem! Thanks!

~AMV

[Tom]

I would use a LENGTH or ATTRIB statement to set the length of a variable.  

The INFORMAT statement is for attaching the INFORMAT, it will only set the length as a side effect.  Now if you do not want SAS to use its default informat behaviour then you could use an INFORMAT statement. Such as if you wanted to read a date or time value.  Of if you wanted to use the $CHAR informat to preserve leading spaces.