R中的条件过滤

Question

我正在尝试为我编写一个自动过滤过程的函数。考虑这个专注于数据数据框的示例：

#generate mock dataframe
set.seed(1)

number <-c("4","5","6")

##groups
colors <- c("red","orange","green")
fruit <- c("apple","tomato","banana")
animal <- c("chicken","pork","cow")


vector_colors <-map(colors, ~ str_c(.,number, sep = "_"))  %>%  reduce(c)
vector_fruit <-map(fruit, ~ str_c(.,number, sep = "_"))  %>%  reduce(c)
vector_animal <-map(animal, ~ str_c(.,number, sep = "_"))  %>%  reduce(c)


vector <- c(vector_colors,vector_fruit,vector_animal)

#dataframe I'll be working with
data <-  tibble(name = vector,
                rank = runif(27, 1, 99))

我想要一个函数或 tidyverse/base R 工作流程，每组打印一个项目（即颜色、水果和动物），以便打印的项目与其组中的所有其他项目相比排名最低。在那个种子示例中，想要的结果将是以下向量：

wanted_result <-c("orange_5","apple_6","cow_6")

Answer 1

制作类型查找表，提取前缀，加入类型，并按类型分组以提取具有最小等级的名称：

library(dplyr)
type_lookup = tibble(
  prefix = c(colors, fruit, animal),
  type = rep(c("color", "fruit", "animal"), times = c(length(colors), length(fruit), length(animal)))
)

data %>%
  mutate(prefix = str_extract(name, pattern = ".*(?=_)")) %>%
  left_join(type_lookup, by = "prefix") %>%
  group_by(type) %>%
  slice_min(rank) %>%
  pull(name)
# [1] "cow_6"    "orange_5" "apple_4"