Subcategories as features

Question

I am using XGbregressor on a dataset with many categorical features. A example looks like this:

shop_name     Category   Sub-category   sub-sub-category   revenue 
              Eateries    Resturants        Sushi
              Eateries    Resturants        Italian
              Eateries     Fast-food        Burger
              Eateries     Fast-food        Burger
              Fashion        Luxury         Jewelery 

   

A)  

          shop_name     Category   Sub-category   sub-sub-category   revenue 
                            1                1             1
                            1                1             2
                            1                2             1
                            1                2             1
                            2                1             1

B)

  shop_name     Category   Sub-category   sub-sub-category   revenue 
                      1                1             1
                      1                1             2
                      1                2             3
                      1                2             3
                      2                3             4

If I just label encode all these features, will Xgboost be able to make out the link between features? And which is a better way to do it, A or B?