If you created your script interactively, you already specified columns in the launch window, so there is no need to specify them again. However, when creating a script from scratch, you normally want to specify the columns to analyze. For example, the following line launches the Distribution platform and specifies the height and weight columns as the Y variables:
Distribution( Y( height,weight ));
Tip: To see all the messages you can send to a column, go to Help > Scripting Index > Data Table > Column Scripting.
Creating a column reference enables you to access and send messages to the column. If you store each column reference in a JSL variable, you can use the variables instead of column names in the Y and X roles and elsewhere in the script.
For example, using Big Class.jmp, the following lines create Ycol as a column reference for weight, and Xcol as a column reference for height. The last line sends a message to Xcol to get the data type.
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );
Ycol = Column( "weight" );
Xcol = Column( "height" );
Xcol << Get Data Type;
In the log, “Numeric” is output for the data type of the Xcol (height column).
Referring to a column reference with the : scoping operator (rather than plain text) is convenient in scripts saved to a data table. If the columns are renamed in the data table, the column references in the script are updated.