我有两个数据框。第一个是引用 df，看起来像这样:

第二个是我的宇宙:

我想合并它们，并只采用宇宙中所有可能对的引用数据框中存在的唯一组合。

理想情况下，我希望报告的数据框看起来像这样:

library(tidyverse)
var1 = c("a","b","c","d")
var2 = c("e","z","f","h")
ref = tibble(var1,var2);ref
sym1 = c("e","b","b","f","n","n","k")
sym2 = c("a","f","z","c","s","k","l")
univ = tibble(sym1,sym2);univ

我如何使用 dplyr 在 R 中执行此操作？

最佳答案

在基础 R 中:

s <- apply(univ, 1, \(x) paste(sort(x), collapse = " "))
r <- paste(ref$var1, ref$var2)
univ[match(r, s), ]

  sym1  sym2 
1 e     a    
2 b     z    
3 f     c    
4 NA    NA   

在一个 tidyverse 友好的管道中:

library(stringr)
library(dplyr)
univ %>% 
  rowwise() %>% 
  mutate(s = str_c(sort(c_across(everything())), collapse = "")) %>% 
  pull(s) %>% 
  match(str_c(ref$var1, ref$var2), .) %>% 
  univ[., ]

https://stackoverflow.com/questions/74233237/