如何将分支/株高/生根/重量等多组特征列分别堆叠后并排整合？

核心思路 #

利用tidyverse中的pivot_longer一次性提取所有特征组的盆号（pot_number）和特征类型，再通过pivot_wider将不同特征展开为并排列，无需多次merge或单独处理单组特征。

代码实现 #

首先构造匹配你数据结构的模拟数据集：

library(tidyverse)

set.seed(123)
df <- tibble(
  treatment = rep(c("A", "B"), each = 3),
  block = rep(c("1", "2", "3"), 2),
  `1branching` = rpois(6, 5),
  `2branching` = rpois(6, 6),
  `3branching` = rpois(6, 4),
  `1_rooting` = rnorm(6, 10, 2),
  `2_rooting` = rnorm(6, 12, 2),
  `3_rooting` = rnorm(6, 9, 2),
  `1height` = rnorm(6, 20, 3),
  `2height` = rnorm(6, 22, 3),
  `3height` = rnorm(6, 19, 3),
  `1weight` = rnorm(6, 5, 1),
  `2weight` = rnorm(6, 6, 1),
  `3weight` = rnorm(6, 4.5, 1)
)

执行数据转换：

df_processed <- df %>%
  # 第一步：将所有特征列转为长格式，提取盆号和特征类型
  pivot_longer(
    cols = -c(treatment, block),  # 排除treatment、block等标识列
    # 正则匹配列名：提取数字（盆号）和特征类型，兼容有无下划线的格式
    names_pattern = "(\\d+)_?(branching|rooting|height|weight)",
    names_to = c("pot_number", "feature"),
    values_to = "value"
  ) %>%
  # 第二步：将特征类型转为列，实现并排展示
  pivot_wider(
    names_from = feature,
    values_from = value
  ) %>%
  # 可选：将盆号转为整数类型
  mutate(pot_number = as.integer(pot_number))

结果说明 #

处理后的数据集结构：

• 保留treatment、block、pot_number标识列
• branching、rooting、height、weight四组特征并排展示，每一行对应一个处理-区组-盆号的完整观测值

适配不同列名格式 #

如果你的列名是branching1、rooting_2这类特征在前、数字在后的格式，只需调整names_pattern和names_to的顺序：

df_processed <- df %>%
  pivot_longer(
    cols = -c(treatment, block),
    names_pattern = "(branching|rooting|height|weight)_?(\\d+)",
    names_to = c("feature", "pot_number"),
    values_to = "value"
  ) %>%
  pivot_wider(names_from = feature, values_from = value) %>%
  mutate(pot_number = as.integer(pot_number))

内容的提问来源于stack exchange，提问作者sofia_0084