It outfits Python with new data types for loading data fast from tabular ... If you want to extract specific rows, you can use a list of the row IDs; for example, df.iloc[[0,1,2,5,7,10,12]].