If columns in .csv file have empty cells, they are identified as "String" automatically when the first 101 cells are empty (or null), and any other value e.g a real value after that cell Spotfire will still consider the data type "String" while it should be "real".
To users, it's inconvenient to adjust hundred of columns in the importing window or checking every single column in the excel file before importing.
Please add the feature where Spotfire keep looking for the first filled cell in the column to detect the column's datatype.
You could also just add a box in the "Advanced" settings to stipulate how many rows to scan and then just "rescan".