求助：在R语言中基于多列值生成合并分类列的实现方法

数据集与需求 #

首先重现你的数据集：

treatment_alone = c("Yes", "No", "Yes", "No", "No", "No", "No", "No")
treatment_and_fertilizer = c("No", "Yes", "No", "Yes", "No", "No", "No", "No")
treatment_and_manure = c("No", "No", "No", "No", "Yes", "No", "Yes", "No")
treatment_and_composte = c("No", "No", "No", "No", "No", "Yes", "No", "Yes")

data_test = data.frame(treatment_alone, treatment_and_fertilizer, treatment_and_manure,
                       treatment_and_composte)

需求：生成一个treatment列，将4个二分类列（Yes/No）合并为单一分类列，每行值对应该行中为"Yes"的列的自定义类别名称（如treatment_alone为Yes时，值为"treatment alone"）。

问题分析 #

你之前的代码存在两个关键问题：

• 第一个代码块的mutate语法错误，未将变量修改逻辑包裹在mutate()括号内；
• coalesce函数仅识别非NA值，你替换后的空字符串""或原有的"No"都不属于NA，因此无法正确匹配目标值。

解决方案 #

以下是三种可行的实现方式：

方法1：dplyr 逐行判断（直观易读） #

使用rowwise()逐行处理，结合case_when匹配每个列的"Yes"状态：

library(dplyr)

data_test %>%
  rowwise() %>%
  mutate(treatment = case_when(
    treatment_alone == "Yes" ~ "treatment alone",
    treatment_and_fertilizer == "Yes" ~ "treatment and fertilizer",
    treatment_and_manure == "Yes" ~ "treatment and manure",
    treatment_and_composte == "Yes" ~ "treatment and compost",
    TRUE ~ NA_character_ # 可选：处理无Yes的行，返回NA
  )) %>%
  ungroup()

方法2：tidyr 转长表处理（适合复杂场景） #

通过宽表转长表，过滤出"Yes"的记录后再转回宽表：

library(dplyr)
library(tidyr)

data_test %>%
  mutate(row_id = row_number()) %>%
  pivot_longer(-row_id, names_to = "treatment_col", values_to = "value") %>%
  filter(value == "Yes") %>%
  mutate(treatment = recode(treatment_col,
    "treatment_alone" = "treatment alone",
    "treatment_and_fertilizer" = "treatment and fertilizer",
    "treatment_and_manure" = "treatment and manure",
    "treatment_and_composte" = "treatment and compost"
  )) %>%
  select(row_id, treatment) %>%
  right_join(data_test %>% mutate(row_id = row_number()), by = "row_id") %>%
  select(-row_id)

方法3：Base R 高效实现（无需加载包） #

利用max.col定位每行"Yes"的位置，再匹配自定义类别名称：

# 定义列名与目标类别的映射关系
treatment_mapping <- c(
  "treatment_alone" = "treatment alone",
  "treatment_and_fertilizer" = "treatment and fertilizer",
  "treatment_and_manure" = "treatment and manure",
  "treatment_and_composte" = "treatment and compost"
)

# 找到每行中值为"Yes"的列索引
yes_col_index <- max.col(data_test == "Yes", ties.method = "first")

# 生成treatment列
data_test$treatment <- treatment_mapping[colnames(data_test)[yes_col_index]]

内容的提问来源于stack exchange，提问作者madina_b