我认为***的方法是在 dplyr

I think the neatest way to do this is in dplyr

library(dplyr)
shop %>% 
  group_by(shop_id, shop_name, city) %>% 
  summarise_all(sum)