| Piotr Krzystek |
Character variables, otherwise known as string variables, contain non-numeric values and are read by a statistical program system as being in text format. The response values for character variables can be discreet and categorical in nature, such as listing single-lettered multiple-choice responses or the names of months. Other character variables can be open-ended, and each observation can possess unique values such as user-entered free text responses to questions in a survey. Some character variables can possess large lengths that may present issues for some statistical programs to read datasets properly. SAS software provides built in solutions to these issues where other statistical programs do not. The purpose of this paper is to present issues that may arise with long character variables and describe how SAS can be utilized to remedy them. The outline for this paper includes defining the types of character variables, common issues that may arise with character variables that are defined as long, and suggestions for SAS code that can be utilized to remedy these issues. The version of SAS utilized for this paper is 9.4, and users would need a basic understanding of data step syntax to edit and run the code. |