根据来自另一个数据帧的一组规则在一个数据帧中创建一组变量答案

【问题标题】：Create a set of variables in a dataframe based on a set of rules from another data frame根据来自另一个数据帧的一组规则在一个数据帧中创建一组变量
【发布时间】：2020-02-10 17:53:07
【问题描述】：

这个问题听起来有点笼统，但我认为举个例子会更清楚：

我有以下两个数据框

数据1

group1     group2       group3     Level 
cat         cat          dog        1
dog         parrot       cat        1
mouse       dolphin      dolphin    1
red         blue         blue       2
green       yellow       green      2
black       purple       cat        2

数据2

var1        level    Score
cat           1        1
dog           1        1
mouse         1        1
dolphin       1        0
parrot        1        1
red           2        1
blue          2        1
green         2        1
purple        2        1
cat           2        0
black         2        0
yellow        2        1

我想修改包含 3 个新列（每个 group1、group2 和 group3 一个）的 data1，其值取决于我在 data2 的“score”列中找到的值，具体取决于“level”的级别（级别是一个因素）。所以基本上我想获得这样的东西：

group1     group2       group3     Level      var1     var2     var3
cat         cat          dog        1          1        1        1
dog         parrot       cat        1          1        1        1
mouse       dolphin      dolphin    1          1        0        0
red         blue         blue       2          1        1        1
green       yellow       green      2          1        1        1
black       purple       cat        2          0        1        0

样本数据

df1 <- structure(list(
  group1 = c("cat", "dog", "mouse", "red", "green", "black"),
  group2 = c("cat", "parrot", "dolphin", "blue", "yellow", "purple"),
  group3 = c("dog", "cat", "dolphin", "blue", "green", "cat"),
  Level = structure(c(1L, 1L, 1L, 2L, 2L, 2L), .Label = c("1", "2"), class = "factor")),
  row.names = c(NA, -6L), class = "data.frame")

df2 <- structure(list(
  var1 = c("cat", "dog", "mouse", "dolphin", "parrot", "red", "blue", "green", "purple", "cat", "black", "yellow"),
  level = structure(c(1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("1", "2"), class = "factor"),
  Score = c(1L, 1L, 1L, 0L, 1L, 1L, 1L, 1L, 1L, 0L, 0L, 1L)),
  row.names = c(NA, -12L), class = "data.frame")

【问题讨论】：

@akrun：好点，我改了
@Darren Tsai：应该的，我无意中写的不一样
@Darren Tsai ：另一个错误，我在 data2 中包含黄色：基本上数据 2 是从 data1 + 手动完成的变量 Score 创建的：输出显示返回表 1 格式匹配的意图每组的变量分数

标签： r

【解决方案1】：

我们可以将第一个数据集转为“长”格式，与第二个数据集连接，然后将其转回“宽”格式

library(dplyr)
library(tidyr)
library(stringr)
df1 %>%
    mutate(rn = row_number()) %>%
    pivot_longer(cols  = -c(rn, Level), values_to = 'var1') %>% 
    rename(level = Level) %>% 
    left_join(df2) %>% 
    mutate(name = str_replace(name, 'group', 'varn')) %>% 
    na.omit %>%
    select(-level, -var1) %>% 
    pivot_wider(names_from = name, values_from = Score, values_fill = list(Score = 0)) %>% 
    select(-rn) %>% 
    bind_cols(df1, .)
#   group1  group2  group3 Level varn1 varn2 varn3
#1    cat     cat     dog     1     1     1     1
#2    dog  parrot     cat     1     1     0     1
#3  mouse dolphin dolphin     1     1     0     0
#4    red    blue    blue     2     1     1     1
#5  green  yellow   green     2     1     0     1
#6  black  purple     cat     2     0     1     0

【讨论】：

【解决方案2】：

我通过purrr::reduce() 递归地将df2 合并到df1 三次。在这一部分中，我将df2 复制了三次，并将它们的第一列名称分别更改为与df1 中的名称匹配。

library(tidyverse)

df2 %>%
  list %>% rep(3) %>%
  imap(~ setNames(.x, c(str_c("group", .y), "Level", str_c("Score", .y)))) %>%
  reduce(left_join, .init = df1)

#   group1  group2  group3 Level Score1 Score2 Score3
# 1    cat     cat     dog     1      1      1      1
# 2    dog  parrot     cat     1      1      1      1
# 3  mouse dolphin dolphin     1      1      0      0
# 4    red    blue    blue     2      1      1      1
# 5  green  yellow   green     2      1      1      1
# 6  black  purple     cat     2      0      1      0

【讨论】：