如何使用tidyverse将dataframe两列转换为方阵格式数据框

核心实现逻辑分三步：

• 提取Factor_1的唯一取值作为方阵的行、列维度基准，过滤掉Factor_2中不属于该集合的冗余因子（比如样例中的d）
• 调用tidyr::pivot_wider完成长表到宽表的结构转换
• 调整列顺序和行顺序匹配，按需将第一列转为行名即可得到目标方阵

完整可复现代码 #

library(tidyverse)

# 构造示例数据
cor_df <- tibble(
  Factor_1 = c("a","a","a","b","b","a","b","b","c","c","c","c"),
  Factor_2 = c("b","a","d","c","b","c","d","a","a","c","d","b"),
  value = c(0.8,1,0.6,0.4,1,0.2,0.75,0.8,0.2,1,0.1,0.4)
)

# 执行转换
target_factors <- unique(cor_df$Factor_1)
cor_matrix <- cor_df %>%
  filter(Factor_2 %in% target_factors) %>%
  pivot_wider(
    id_cols = Factor_1,
    names_from = Factor_2,
    values_from = value
  ) %>%
  select(Factor_1, all_of(target_factors)) %>%
  column_to_rownames(var = "Factor_1")

输出结果 #

运行后cor_matrix的输出和目标效果完全一致：

a   b   c
a 1.0 0.8 0.2
b 0.8 1.0 0.4
c 0.2 0.4 1.0

补充说明 #

• 原始数据中已经包含因子自相关值为1的记录，无需额外手动填充对角线
• 如果后续遇到相关系数配对缺失（比如仅存储了a-b的相关值，未存储b-a的对应值），可以在过滤步骤后增加对称补全逻辑，当前样例数据配对完整可直接运行
• 如果需要保留tibble格式，删除最后一步column_to_rownames即可，第一列会保留因子名称。

内容的提问来源于stack exchange，提问作者JEJI