The first wizard page give you the option of basing the reader on an existing data file, or defining the data source from scratch.
If you selected a to infer the structure from an existing file then the next page allows you to define the field delimiters.