如何在 R 中同时使用计数函数和 case_when？

Question

我有一个数据框，我想计算每个名称符合特定条件的观察次数：

library(dplyr)
test <- tibble(name = c("Justin", "Corey", "Sibley", "Justin", "Corey", "Sibley", "Justin", "Corey", "Sibley"),
               class = c("Bio", "Bio", "Bio", "Psych", "Pysch", "Psych", "English", "English", "English"),
               result = c("Fail", "Pass", "Pass", "Fail", "Pass", "Pass", "Fail", "Fail", "Pass"))

在上面的示例中，学生学习的课程是 STEM（例如“Bio”和“Psych”）或文学（“English”）。我想创建两列新列，一列表示他们的 STEM 分数，另一列包含他们的文学分数，其中代码每通过一次计为 1。

答案应该是这样的：

library(dplyr)
answer <- tibble(name = c("Justin", "Corey", "Sibley"),
                 stem_assessment = c(0, 2, 2),
                 lit_assessment = c(0, 0, 1))

我尝试过使用case_when()、group_by()、n()、summarize() 和count()，但我似乎无法破解它。

Answer 1

group_by 和 summarise 应该这样做 -

library(dplyr)

stem <- c("Bio", "Pysch", "Psych")
lit <- c('English')

test %>%
  group_by(name) %>%
  summarise(stem_assessment = sum(class %in% stem & result == "Pass"),
            lit_assessment = sum(class %in% lit & result == "Pass"))

#   name   stem_assessment lit_assessment
#  <chr>            <int>          <int>
#1 Corey                2              0
#2 Justin               0              0
#3 Sibley               2              1

Answer 2

您可以使用count() 和pivot_longer/pivot_wider 的组合。我还在test 中编码了错字（Pysch）。

library(tidyverse)

test %>%
  mutate(
    stem_assessment = class %in% c("Bio", "Psych", "Pysch") & result == "Pass",
    lit_assessment = class == "English" & result == "Pass"
  ) %>% 
  select(-c(class, result)) %>%
  pivot_longer(-name, names_to = "var", values_to = "value") %>% 
  count(name, var, wt = value) %>% 
  pivot_wider(name, names_from = var, values_from = n)