Is there a way to group within groups?

I have data that is divided into major categories, followed by a subcategories with a subcategory within those. For example:

2	Panera Bread
3	Sandwiches
4	Turkey
4	Roast Beef
4	Ham
3	Soups
4	French Onion
4	Broccoli Cheddar
4	Black Bean
4	Chicken Noodle
2	McDonalds
3	Sandwiches
4	Hamburger
4	Chicken
4	Fish
3	Milkshakes
4	Chocolate
4	Vanilla
4	Strawberry
4	Shamrock

I want to group all of the rows with "4" in column A, and then keep those grouped while grouping all of the rows with "3" in Column A, so that when everything is collapsed, I'm only left with a list of "2"s.

Is this possible? Whenever I try to group the "4"s and "3"s, it ends up combining them as one big group instead of doing the tiers that I want.

Could someone try giving me an example of how to do this with the rows between "Panera Bread" and "McDonalds"?

Sorry for my silly example

No, I want to keep the subcategories under the correct main category - I believe autofiltering would just put them in number order?

You can only group consecutive rows/columns.

Here's an alternative idea.

Double click on Col.A will hide all the rows have same value.
Double click on Col.B will unhide all rows.