By doing so, you can have one data file for many people and/or project files and use the data relevant to the specific scenario – leaving out the rest.
On large datasets, the possibility of importing only a small portion of the columns will greatly improve the overall performance of the system.