如何合并数据帧并根据其他行的值同时重新编码 NA？

Question

我想做一些应该相对容易的事情，但我发现这是不可能的。 （在 R 中相对较新）。

我有这两个数据集：

df1 <- data.frame("Basket" = c("W1", "W1"),          
             "Fruits" = c("apple", "banana"),
             "Color" = c("green", "yellow"),
             "Spanishfruit" = c("manzana", "platano"),
             "Spanishcolor" = c("verde", "amarillo"))

df <- data.frame("Basket" = c("W2", "W2", "W2"),          
              "Fruits" = c("apple", "banana", "grape"),
              "Color" = c("green", "yellow", "purple"),
              "Spanishfruit" = c(NA, NA, NA),
              "Spanishcolor" = c(NA, NA, NA))
             

首先，我想合并它们（我为此使用了 bind_rows），这是可行的。但是，我还想在第二个数据框中重新编码 NA，以便如果数据集之间的变量 fruit 和 color 相等（而变量篮子不相等），变量“Spanishfruit”和“Spanishcolour”（简单地翻译成西班牙语）是平等的。即，我只设法做到了这一点：

df3 <- data.frame("Round" = c("W1", "W1","W2", "W2", "W2" ),          
              "Fruits" = c("apple", "banana", "apple", "banana", "grape"),
              "Color" = c("green", "yellow", "green", "yellow", "purple"),
              "Spanishfruit" = c("manzana", "platano", NA, NA, NA),
              "Spanishcolor" = c("verde", "amarillo", NA, NA, NA)) 

但我想要的是这个。

df4 <- data.frame("Round" = c("W1", "W1","W2", "W2", "W2" ),          
              "Fruits" = c("apple", "banana", "apple", "banana", "grape"),
              "Color" = c("green", "yellow", "green", "yellow", "purple"),
              "Spanishfruit" = c("manzana", "platano", "manzana", "platano", NA),
              "Spanishcolor" = c("verde", "amarillo", "verde", "amarillo", NA))

到目前为止，我尝试了几件事，主要是结合了 mutate 和 case_when。我认为解决方案可能是在合并它们时合并和建立正确的键？我真的迷路了，无法在任何地方找到解决这个问题的方法。我想要的是，如果“水果”和“颜色”列中的值在 df1 的一行和 df2 的另一行中相等，当我合并它们时，来自 df2 的 rwo 中的“Spanishfruit”和“Spanishcolor”中的值等于 df1 中的行中的值，而变量篮子仍然不同。或者，换一种说法：如果跨行的两个变量的值相等，我希望两个变量的值也相等，同时保持另一个变量不同。

如果有人知道如何解决这个问题，我将非常感激。找到一个基于 tidyverse 的解决方案会更有帮助，但一切都会有所帮助。

谢谢！

Answer 1

这是您要查找的输出吗？

library(tidyverse)

df1 <- data.frame("Basket" = c("W1", "W1"),          
               "Fruits" = c("apple", "banana"),
               "Color" = c("green", "yellow"),
               "Spanishfruit" = c("manzana", "platano"),
               "Spanishcolor" = c("verde", "amarillo"))

df <- data.frame("Basket" = c("W2", "W2", "W2"),          
              "Fruits" = c("apple", "banana", "grape"),
              "Color" = c("green", "yellow", "purple"),
              "Spanishfruit" = c(NA, NA, NA),
              "Spanishcolor" = c(NA, NA, NA))

df3 <-rbind(df, df1)

df3 %>% group_by(Fruits) %>%
        mutate(Spanishfruit = zoo::na.locf(Spanishfruit, na.rm = FALSE, fromLast = TRUE),
               Spanishcolor = zoo::na.locf(Spanishcolor, na.rm = FALSE, fromLast = TRUE)) %>%
       arrange(Basket, desc())
#> # A tibble: 5 x 5
#> # Groups:   Fruits [3]
#>   Basket Fruits Color  Spanishfruit Spanishcolor
#>   <fct>  <fct>  <fct>  <chr>        <chr>       
#> 1 W2     apple  green  manzana      verde       
#> 2 W2     banana yellow platano      amarillo    
#> 3 W2     grape  purple <NA>         <NA>        
#> 4 W1     apple  green  manzana      verde       
#> 5 W1     banana yellow platano      amarillo

^{由reprex package (v0.3.0) 于 2021-01-07 创建}