r - 查找数据框中字符串的百分比，但每行只计算一次

Question

我的数据框看起来像这样:

id fruit1 fruit2 fruit3 
1  apple  banana orange
2  banana
3  apple  apple
4  banana apple
5  orange apple

有没有办法找到每个分母为 5 的字符串所占个体的百分比？

因此结果将是苹果 = .8、香蕉 = .6 和橙子 = .4

我正在使用的实际数据库很大，因此如果解决方案不需要键入每个字符串，那就太好了。

Answer 1

tidyverse 中的一个选项将被 reshape 为“长”格式，然后旋转回“宽”并获取平均值

library(dplyr)
library(tidyr)
dd %>% 
  # // reshape to long format
  pivot_longer(cols = -id) %>%
  # // remove the blank rows
  filter(value != '') %>%
  # // get the distinct rows
  distinct(id, value) %>% 
  # // reshape to wide format
  pivot_wider(names_from = value, values_from = value, 
      values_fn = list(value = length), values_fill = list(value = 0)) %>% 
  # get the mean of columns
  summarise(across(apple:orange, mean))
# A tibble: 1 x 3
#  apple banana orange
#  <dbl>  <dbl>  <dbl>
#1   0.8    0.6    0.4

数据

dd <- structure(list(id = 1:5, fruit1 = c("apple", "banana", "apple", 
"banana", "orange"), fruit2 = c("banana", "", "apple", "apple", 
"apple"), fruit3 = c("orange", "", "", "", "")),
class = "data.frame", row.names = c(NA, 
-5L))