In a structured SAS table, each column must have three required attributes: name, length, and type. SAS column names can be 1 to 32 characters long. The name must start with a letter or underscore, and can continue with any combination of numbers, letters, or underscores. SAS column names can be uppercase, lowercase, or mixed case. Depending on which environment you use to submit your SAS code, SAS might allow for spaces and special symbols other than underscores in column and table names. If you use data sources other than SAS that have flexible column name rules, SAS can make allowances for that. However, for simplicity and consistency, we really recommend following these SAS naming conventions. Table names should follow the same rules as well. The column type can be either character or numeric. SAS Studio helps us identify column type with either the A or 1, 2, 3 icons. Numeric columns can store only numeric values, which can include the digits 0 through 9, a minus sign, a single decimal point, an E for scientific notation. Character columns can store letters, numbers, special characters, and blanks, basically anything you can type. Let's take a moment to talk about dates in SAS. I'll open a different data source that has a column named RawDate. The column is numeric and represents the number of days from January 1, 1960. These numeric dates makes sorting and calculations easy, but they obviously don't look like dates that we can understand. There are dozens of different SAS date formats we can apply to improve how the raw values appear. I'll create a copy of RawDate as FormattedDate, just to see the columns side by side. I'll use a FORMAT statement to apply the DATE9. format. After running this step, we can see that FormattedDate shows up as day-month-year. Or I can go back to the program and change the format to WORDDATE. And now we see the full month, name, day, and year. I'll go back to the other program to continue discussing column attributes. Length is the last required attribute, and it represents the number of bytes allocated to store column values. The length is related to the column type. Notice that each of the numeric columns have a length of 8. This is the default, which is enough to store about 16 significant digits. Character columns can be any length between 1 and 32,767 bytes, and a byte stores one character. If we look at Region, the length is 25, which is sufficient for the longest region name, Central America/Caribbean. The length of subsidiary is only 12. Optional attributes that we have discussed in our courses include labels and formats. If we look at the Profit column, notice there is a label that was assigned in the DATA step code. And the $12. format is assigned, which displays numbers with a dollar sign and commas rounded to the nearest whole number.
