如何将含列表列的DataFrame重塑为指定结构（R语言）

现有如下结构的DataFrame，希望通过unnest操作将files列展开为以Files.为前缀的新列，得到指定的df2结构。当前尝试的代码无法实现目标，请问该如何操作？

原始代码 #

library(dplyr)
library(tidyr)
library(stringr) # 原代码用到str_c但未加载该包

df <-  structure(list(url = "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", 
                       files = "list(sha = c(\"bca42f706ba9ef5b822cfc8581ca79bec93f28c2\", \"9f29ababe866e9d6df73bb730a0eb5d216562838\", \"3c3fe578da77c9d2dc92e339ecd4d954ed24ba15\", \"03f52efb0d0be91e580dd1d636b5fcc54033aacc\", \"6fd0a99f12295e0ce9ea56062ced4a92a7152470\", \"abfeef07a423b9fcae17e4a08de9a93f39448c25\", \"abfeef07a423b9fcae17e4a08de9a93f39448c25\", \"f98b02b85225bdf54206e1955aaa34ebe298899e\", \"d97d3cf93f85d48dd09dba15f7b0f652429980c3\"), filename = c(\"src/manifest.json\", \"src/skin/.DS_Store\", \"src/skin/dashboard.html\", \"src/skin/dist/main.dev.js\", \"src/skin/dist/welcome.dev.js\", \"src/skin/images/icon.png\", \"src/skin/images/icon2.png\", \"src/skin/main.js\", \"src/skin/welcome.html\"), status = c(\"modified\", \"modified\", \"modified\", \"removed\", \"removed\", \"modified\", \"removed\", \"modified\", \"modified\"), additions = c(8, 0, 0, 0, 0, 0, 0, 0, 4), deletions = c(8, 0, 1, 203, 142, 0, 0, 3, 3), changes = c(16, 0, 1, 203, 142, 0, 0, 3, 7))"), row.names = 1L, class = "data.frame")
                       
df_reshaped <- df %>%
  rename_all(funs(str_c("Files.", .))) %>%
  unnest(files)

期望结果 #

df2 <- data.frame(url = c("https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702","https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702", "https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702","https://api.github.com/repos/aleecia/CCPABrowserTool/commits/fc795a994d6284633393ca940796ce82fd69f702"),
           Files.sha = c("bca42f706ba9ef5b822cfc8581ca79bec93f28c2", "9f29ababe866e9d6df73bb730a0eb5d216562838", "3c3fe578da77c9d2dc92e339ecd4d954ed24ba15", "03f52efb0d0be91e580dd1d636b5fcc54033aacc", "6fd0a99f12295e0ce9ea56062ced4a92a7152470", "abfeef07a423b9fcae17e4a08de9a93f39448c25", "abfeef07a423b9fcae17e4a08de9a93f39448c25", "f98b02b85225bdf54206e1955aaa34ebe298899e", "d97d3cf93f85d48dd09dba15f7b0f652429980c3"),
           Files.filename = c("src/manifest.json", "src/skin/.DS_Store", "src/skin/dashboard.html", "src/skin/dist/main.dev.js", "src/skin/dist/welcome.dev.js", "src/skin/images/icon.png", "src/skin/images/icon2.png", "src/skin/main.js", "src/skin/welcome.html"),
           Files.status = c("modified", "modified", "modified", "removed", "removed", "modified", "removed", "modified", "modified"),
           Files.additions = c(8, 0, 0, 0, 0, 0, 0, 0, 4),
           Files.deletions = c(8, 0, 1, 203, 142, 0, 0, 3, 3),
           Files.changes = c(16, 0, 1, 203, 142, 0, 0, 3, 7))

解决方案 #

你的代码存在两个核心问题：

1. files列是字符串格式，不是可直接展开的列表对象，unnest无法识别；
2. 提前用rename_all给所有列加前缀，会导致url列也被重命名，不符合预期。

正确步骤如下：

1. 先把files列的字符串转换成实际的列表结构；
2. 展开files列得到子列；
3. 给展开后的子列添加Files.前缀，保留原url列不变。

完整实现代码：

library(dplyr)
library(tidyr)
library(stringr)

df_reshaped <- df %>%
  # 将字符串格式的files转换为可操作的列表
  mutate(files = lapply(files, function(x) eval(parse(text = x)))) %>%
  # 展开files列
  unnest(files) %>%
  # 给除url外的所有列添加Files.前缀
  rename_with(~str_c("Files.", .), -url)

运行后即可得到与df2结构一致的结果。

内容的提问来源于stack exchange，提问作者foc