如何在 R 中为多个变量创建计数列

Question

我有一个看起来像这样的数据集。每个事件都有 1-3 个标签，其中列出了颜色。所有颜色都可以出现在 3 个标签列中的任何一个中。一项可以有多个事件。

> data.frame(item = c(1,1,2,3,4,4,4,5), event = seq(1,8), 
+            tag1 = c("red","red","blue","green","red","yellow","black","purple"), 
+            tag2 = c("blue","NA","NA","yellow","orange","NA","purple","red"), 
+            tag3 = c("NA","NA","NA","red","magenta","NA","red","green"))
  item event   tag1   tag2    tag3
1    1     1    red   blue      NA
2    1     2    red     NA      NA
3    2     3   blue     NA      NA
4    3     4  green yellow     red
5    4     5    red orange magenta
6    4     6 yellow     NA      NA
7    4     7  black purple     red
8    5     8 purple    red   green

我想要做的是转换数据，以便每个项目都有一行，然后每个可能的颜色都有一列，其值是该项目是否曾经被标记过。

例如：

| Item| Red | Blue | Green | etc |
| 1   | 1   | 1    | 0     | 0   |
| 2   | 0   | 0    | 1     | 0   | 

我不想手动创建列，因为在我的实际代码中，大约有 800 种不同的“颜色”。

我正在与 tidyverse 合作尝试解决此问题，但我对其他使其更容易的软件包持开放态度。

运行时间确实很重要，因为数据集中有超过 1000 万个事件。

Answer 1

library(tidyverse)\n\ndf %>%\n  na_if("NA") %>% \n  pivot_longer(cols = starts_with("tag")) %>% \n  drop_na() %>% \n  count(item, value) %>% \n  pivot_wider(names_from = value, \n              values_from = n, \n              values_fill = 0)\n\n# A tibble: 5 \xc3\x97 9\n   item  blue   red green yellow black magenta orange purple\n  <dbl> <int> <int> <int>  <int> <int>   <int>  <int>  <int>\n1     1     1     2     0      0     0       0      0      0\n2     2     1     0     0      0     0       0      0      0\n3     3     0     1     1      1     0       0      0      0\n4     4     0     2     0      1     1       1      1      1\n5     5     0     1     1      0     0       0      0      1\n

\n

要求：

\n

df %>%\n  na_if("NA") %>% \n  pivot_longer(cols = starts_with("tag")) %>% \n  drop_na() %>% \n  count(item, value) %>% \n  pivot_wider(names_from = value, \n              values_from = n, \n              values_fill = 0) %>% \n  mutate(grey = 0, \n         white = 0)\n

\n