如何在R中将两个melt/gather/pivot_longer命令合并为单个命令？

错误原因说明 #

你遇到的问题是因为**reshape2::melt 不支持 data.table::patterns 参数**——patterns 是data.table包中melt函数的专属语法，混用两个包的函数会导致找不到对应参数或匹配失败的错误。

方案1：使用data.table的melt（高效原生支持多组列） #

data.table的melt函数原生支持多组列的批量重塑，直接通过patterns匹配列名前缀即可：

library(data.table)

# 将data.frame转为data.table
setDT(df)

# 一次完成多组列的宽转长
long_df <- melt(df,
                id.vars = "id",  # 保留的标识符列
                measure.vars = patterns(time = "^time", count = "^count"),  # 匹配两组列的前缀
                variable.name = c("group_num"),  # 提取的分组编号（1-4）
                value.name = c("time_count_value", "count_value")  # 对应值的列名
)

# 将分组编号转为要求的timepoint/count_no格式
long_df[, `:=`(
  timepoint = paste0("time", group_num),
  count_no = paste0("count", group_num)
)][, group_num := NULL]  # 删除临时分组列

# 调整列顺序以匹配目标格式
setcolorder(long_df, c("timepoint", "time_count_value", "count_no", "count_value", "id"))

方案2：使用tidyr的pivot_longer（tidyverse风格，语法直观） #

利用tidyr的pivot_longer结合正则表达式提取列名中的前缀和编号，一次完成重塑：

library(tidyr)
library(dplyr)

long_df <- df %>%
  # 匹配所有time和count开头的列，提取前缀和数字编号
  pivot_longer(
    cols = starts_with(c("time", "count")),
    names_to = c(".value", "group_num"),
    names_pattern = "(time|count)(\\d)"  # 正则捕获：前缀（time/count）+ 数字
  ) %>%
  # 生成要求的timepoint和count_no列
  mutate(
    timepoint = paste0("time", group_num),
    count_no = paste0("count", group_num)
  ) %>%
  # 调整列顺序并删除临时分组列
  select(timepoint, time_count_value = time, count_no, count_value = count, id) %>%
  arrange(id, group_num) %>%
  select(-group_num)

方案3：reshape2分两次合并（兼容旧版工具链） #

如果必须使用reshape2，可以分别重塑time和count列，再通过分组编号合并：

library(reshape2)

# 重塑time列
time_long <- melt(df,
                  id.vars = "id",
                  measure.vars = starts_with("time"),
                  variable.name = "timepoint",
                  value.name = "time_count_value") %>%
  mutate(group_num = gsub("time", "", timepoint))

# 重塑count列
count_long <- melt(df,
                   id.vars = "id",
                   measure.vars = starts_with("count"),
                   variable.name = "count_no",
                   value.name = "count_value") %>%
  mutate(group_num = gsub("count", "", count_no))

# 合并并整理格式
long_df <- merge(time_long, count_long, by = c("id", "group_num")) %>%
  select(timepoint, time_count_value, count_no, count_value, id) %>%
  arrange(id, group_num) %>%
  select(-group_num)

内容的提问来源于stack exchange，提问作者Feli