根据条件更新多行

Question

我想在一列的基础上同时更新三列

我的数据是这样的

df <- data.frame(input = c("Antidesma cuspidatum Mull.Arg.", "Antidesma cuspidatum Müll.Arg.", 
                  "Alchornea parviflora (Benth.) Mull.Arg.", "Alchornea parviflora (Benth.) Müll.Arg."),
                 n1 = c("Antidesma cuspidatum", NA, "Alchornea parviflora", NA),
                 n2 = c("Antidesma", NA, "Alchornea", NA),
                 n3 = c("Phyllanthaceae", NA, "Euphorbiaceae", NA))

                                    input                   n1        n2             n3
1          Antidesma cuspidatum Mull.Arg. Antidesma cuspidatum Antidesma Phyllanthaceae
2          Antidesma cuspidatum Müll.Arg.                 <NA>      <NA>           <NA>
3 Alchornea parviflora (Benth.) Mull.Arg. Alchornea parviflora Alchornea  Euphorbiaceae
4 Alchornea parviflora (Benth.) Müll.Arg.                 <NA>      <NA>           <NA>

请问如果我发现 input 列的前两个 strings 相同，则相应的行也相同。这意味着本例中 n1、n2、n3 的值（第 2 行和第 4 行）将加上值（第 1 行和第 3 行）。

这里是我想要的输出

                                    input                   n1        n2             n3
1          Antidesma cuspidatum Mull.Arg. Antidesma cuspidatum Antidesma Phyllanthaceae
2          Antidesma cuspidatum Müll.Arg. Antidesma cuspidatum Antidesma Phyllanthaceae
3 Alchornea parviflora (Benth.) Mull.Arg. Alchornea parviflora Alchornea  Euphorbiaceae
4 Alchornea parviflora (Benth.) Müll.Arg. Alchornea parviflora Alchornea  Euphorbiaceae

这个案例对我有什么建议吗？

Answer 1

您可以使用 dplyr 包。首先，我创建一个列 gr，其中仅包含 input 的前两个字符串。然后我通过将该组的非 NA 值放在那里来更改（或 mutate）列 n1、n2 和 n3。

library(dplyr)

df %>%
  group_by(gr = gsub("(^\w+ \w+) .*", "\1", input)) %>%
  mutate(across(c(n1, n2, n3), ~.x[!is.na(.x)][1])) %>%
  ungroup()

Answer 2

基础 R 解决方案：

# Resolve the names of column vectors prefixed with "n":
# na_col_names => character vector
na_col_names <- grep(
  "n\d+",
  names(df),
  value = TRUE
)

# Carry the last value forward: df => data.frame
df[,na_col_names] <- lapply(
  na_col_names,
  function(x){
    df[,x] <- na.omit(df[,x])[cumsum(!(is.na(df[,x])))]  
  }
)

Tidyverse:

library(tidyverse)
df %>% 
  mutate_if(
    str_detect("n\d+", names(.)),
    function(x){
      fill(x, .direction = "down")
    }
  )

根据条件更新多行

update multiple rows based on conditions

r

regex

dplyr

tidyverse

stringr