另一个维护行排序的tidyverse方法：

df %>%
  group_by(Name)%>%
  mutate(Category = ifelse("Special"%in%Category ,'Special', 
                           Category[which.max(Number)]))
# A tibble: 7 × 3
# Groups:   Name [3]
  Name   Category   Number
  <chr>  <chr>       <int>
1 Name A Category 1     10
2 Name A Category 1      5
3 Name B Category 2      3
4 Name B Category 2     20
5 Name C Special         3
6 Name C Special        25
7 Name C Special         4

我们可以这样做：
通过以降序形式排列（-），我们保持了在mutate中使用first()函数的行顺序：

library(dplyr)

df %>% 
  group_by(Name) %>% 
  arrange(-Number, .by_group = TRUE) %>% 
  mutate(Category = ifelse("Special" %in% Category, "Special", first(Category)))

Name   Category   Number
  <chr>  <chr>       <int>
1 Name A Category 1     10
2 Name A Category 1      5
3 Name B Category 2     20
4 Name B Category 2      3
5 Name C Special        25
6 Name C Special         4
7 Name C Special         3

df <- structure(list(Name = c("Name A", "Name A", "Name B", "Name B", 
"Name C", "Name C", "Name C"), Category = c("Category 1", "Category 2", 
"Category 1", "Category 2", "Category 1", "Category 2", "Special"
), Number = c(10L, 5L, 3L, 20L, 3L, 25L, 4L)), class = "data.frame", row.names = c(NA, 
-7L))