I have a data set that collects information from a process. There are cases when the information is not typed into the collection interface that results in blanks for that data column. I would like to fill those blanks with "NA" so they appear in the distribution analysis. Is there a way to do this without creating a whole new column?