如何将数据框重塑为长格式并去除NA值？解决reshape报错问题

错误原因分析 #

你遇到的'varying' arguments must be the same length错误，核心问题有两个：

• reshape()的varying参数中，两个列表长度不匹配：第一个列表包含4个Researcher变量，第二个只有3个revisit变量
• 变量名书写错误：实际数据里是Researcher1/revisit1（无空格），但代码里写成了Researcher 1/revisit 1

正确解决方案 #

推荐用tidyr包的pivot_longer函数，比base R的reshape更直观，能直接生成无NA的长格式：

步骤1：加载包并重现数据 #

library(tidyr)
library(dplyr)

# 还原你的原始数据
survey <- data.frame(
  zone = c(3,3,1,1,2,2,4,4),
  situation = c("Answered","Refusal","Answered","Refusal","Answered","Refusal","Answered","Refusal"),
  Researcher1 = c(NA,NA,10,7,NA,NA,NA,NA),
  Researcher2 = c(NA,NA,NA,NA,15,3,NA,NA),
  Researcher3 = c(4,15,NA,NA,NA,NA,NA,NA),
  Researcher4 = c(NA,NA,NA,NA,NA,NA,13,8),
  revisit1 = c(NA,NA,2,5,3,0,NA,NA),
  revisit2 = c(10,5,NA,NA,NA,NA,NA,NA),
  revisit3 = c(NA,NA,NA,NA,NA,NA,4,4)
)

步骤2：重塑为无NA的长格式 #

long_rsch <- survey %>%
  # 转换Researcher系列列为长格式，自动丢弃NA
  pivot_longer(cols = starts_with("Researcher"), 
               names_to = "team", 
               values_to = "first_team",
               values_drop_na = TRUE) %>%
  # 转换revisit系列列为长格式，自动丢弃NA
  pivot_longer(cols = starts_with("revisit"), 
               names_to = "revisit_team", 
               values_to = "Revisit",
               values_drop_na = TRUE) %>%
  # 清理team名称（去掉前缀，只保留数字）
  mutate(team = gsub("Researcher", "", team),
         revisit_team = gsub("revisit", "", revisit_team)) %>%
  # 按分区和应答情况排序
  arrange(zone, situation)

最终输出结果 #

运行后得到的结果与你的期望一致：

若坚持使用base R的reshape函数 #

需要补全varying列表长度，同时修正变量名：

long_rsch_base <- reshape(survey, 
                          direction = "long", 
                          idvar = c("zone", "situation"),
                          varying = list(c("Researcher1","Researcher2","Researcher3","Researcher4"),
                                         c("revisit1","revisit2","revisit3",NA)), # 补NA让两个列表长度一致
                          v.names = c("first_team", "Revisit"),
                          times = c(1,2,3,4),
                          timevar = "team") %>%
  na.omit() %>%
  arrange(zone, situation)

内容的提问来源于stack exchange，提问作者Kledson Lemes