如何在R语言中根据指定列名从不同列取值生成新列？

嘿，我来帮你搞定这个需求！你想要在数据框df里新增一列val，让它根据每行day列给出的列名，从对应列中提取值对吧？你的思路方向没问题，但之前写的函数在索引处理上有偏差，我给你几个简单靠谱的解决方案：

先构造示例数据 #

首先咱们先把你给出的示例数据框准确构造出来，方便测试：

df <- data.frame(
  id = c(10,12,13,14,16,19,20,21,23),
  day = c("wed","wed","tue","thu","thu","fri","sun","mon","mon"),
  mon = c("a","b","h","j","y","e","y","u","i"),
  tue = c("y","e","y","u","i","y","e","y","u"),
  wed = c("b","h","j","b","h","j","y","a","b"),
  thu = c("j","y","b","h","j","y","a","b","h"),
  fri = c("j","b","h","j","y","a","b","h","j"),
  sat = c("b","h","j","b","h","j","y","a","b"),
  sun = c("a","b","h","j","y","e","y","u","i"),
  stringsAsFactors = FALSE # 避免字符自动转因子带来的问题
)

方案一：基础R原生方法（高效简洁） #

用矩阵索引的方式可以快速实现需求，这是最原生的R写法，不需要额外包：

# 生成val列
df$val <- df[cbind(1:nrow(df), match(df$day, colnames(df)))]

原理解释： #

• match(df$day, colnames(df))：把每行的day值转换成它在数据框列名中对应的位置索引（比如"wed"对应的列索引是4）
• cbind(1:nrow(df), ...)：把行索引（从1到数据框总行数）和上面得到的列索引组合成一个矩阵，R会用这个矩阵来精准提取每个位置的对应值

方案二：tidyverse/dplyr风格写法 #

如果你习惯用dplyr进行数据处理，也可以用行分组的方式实现：

library(dplyr)

df <- df %>%
  rowwise() %>% # 让数据按行单独处理
  mutate(val = pull(cur_data(), day)) %>% # 从当前行提取day指定列的值
  ungroup() # 取消行分组，恢复常规数据框结构

验证结果 #

运行完上面任意一种方法后，你可以打印df查看结果，比如前几行的val列应该是：第一行b（对应wed列）、第二行h（对应wed列）、第三行y（对应tue列），完全符合你的预期。

关于你之前函数的问题 #

你之前写的lookupfunction里，which(x=x)这部分逻辑不对——x是整个df$day向量，这样没法拿到每行对应的行号，而且函数最后也没有返回对应列的取值，所以没法生成正确的val列。上面的两种方法都比自定义函数更高效简洁哦～

内容的提问来源于stack exchange，提问作者Andreas Hoff