将数据框转换回嵌套列表答案

【问题标题】：Convert dataframe back to nested list将数据框转换回嵌套列表
【发布时间】：2020-11-17 01:34:02
【问题描述】：

我正在尝试在 YAML、列表和数据框之间进行转换。我有一个如下所示的 YAML：

sheet1:
  var1:
    type: character
    description: description for my var1
  var2:
    type: character
    description: description for my var2
sheet2:
  var3:
    type: character
    description: description for my var3
  var4:
    type: character
    description: description for my var4

我正在阅读它并使用以下脚本转换为数据框：

library(yaml)
library(tidyverse)
yaml_file <- yaml::read_yaml("yaml.yaml")
mydf <- yaml_file %>% map(bind_rows) %>% bind_rows(.id = "sheet")

在某些情况下，我需要将数据帧转换回我原来的 YAML。以下是我目前不成功的尝试。

a <- split(mydf, mydf$sheet)
b <- lapply(a, function(x) split(x, f = x[["var"]]))
c <- rapply(b, as.list, how = "list")
cat(as.yaml(c)) # Not the same as my input

【问题讨论】：

您可能无法回到原来的状态，因为在绑定列表时，您会丢弃名称
小评：data.table和tidyverse语法最好不要混用； tidyverse 方式将是 mydf <- yaml_file %>% map(bind_rows) %>% bind_rows(.id = "sheet")。
@RonakShah，输出如下：list(sheet1 = list(var1 = list(type = "character", description = "description for my var1"), var2 = list(type = "character", description = "description for my var2")), sheet2 = list(var3 = list(type = "character", description = "description for my var3"), var4 = list(type = "character", description = "description for my var4")))

标签： r yaml

【解决方案1】：

首先正如@Mauritus 所说，让我们要么维护tidyverse，要么只维护data.table。

这是tidyverse的一种方式：

library(tidyverse)
yaml_file <- yaml::read_yaml("yaml.yaml")
mydf <- yaml_file %>% 
     map(bind_rows, .id='var') %>% 
     bind_rows(.id = "sheet")
mydf

# A tibble: 4 x 4
  sheet  var   type      description            
  <chr>  <chr> <chr>     <chr>                  
1 sheet1 var1  character description for my var1
2 sheet1 var2  character description for my var2
3 sheet2 var3  character description for my var3
4 sheet2 var4  character description for my var4

这样可以确保没有数据丢失。在您的情况下，您丢失了 var 部分：现在将其组合回yaml 文件：

 map(split(mydf[-1], mydf$sheet),~split(.x[-1], .x$var))%>%
     yaml::as.yaml()%>%
     cat()

sheet1:
  var1:
    type: character
    description: description for my var1
  var2:
    type: character
    description: description for my var2
sheet2:
  var3:
    type: character
    description: description for my var3
  var4:
    type: character
    description: description for my var4

【讨论】：

短：yaml_file %>% map_df(~bind_rows(.x, .id = 'var'), .id = 'sheet')