如何生成层级JSON：让叶子节点挂载至最近非空父节点

核心思路 #

将数据拆分为Level3非空和为空的两组，分别执行不同嵌套逻辑：非空组按Level1→Level2→Level3层级挂载叶子节点；为空组跳过Level3，直接将叶子节点挂载到最近的非空父节点（Level2），最后合并两组数据再转换为JSON。

具体代码实现 #

假设你的原始数据框名为df，包含Level1、Level2、Level3、Name、Id列，示例代码如下：

library(dplyr)
library(tidyr)
library(jsonlite)

# 示例测试数据
df <- tibble(
  Level1 = c("A", "A", "B", "B"),
  Level2 = c("A1", "A2", "B1", "B2"),
  Level3 = c("A1a", NA, "B1a", NA),
  Name = c("Leaf1", "Leaf2", "Leaf3", "Leaf4"),
  Id = c(1, 2, 3, 4)
)

# 处理Level3非NA的情况：完整三级嵌套
df_non_na <- df %>%
  filter(!is.na(Level3)) %>%
  nest(children = c(Name, Id)) %>%
  nest(children = c(Level3, children)) %>%
  nest(children = c(Level2, children))

# 处理Level3为NA的情况：跳过Level3，直接挂载到Level2
df_na <- df %>%
  filter(is.na(Level3)) %>%
  select(-Level3) %>%
  nest(children = c(Name, Id)) %>%
  nest(children = c(Level2, children))

# 合并两组数据，按Level1统一整合子节点
final_df <- bind_rows(df_non_na, df_na) %>%
  group_by(Level1) %>%
  summarise(children = list(bind_rows(children))) %>%
  ungroup()

# 转换为符合要求的JSON
final_json <- toJSON(final_df, pretty = TRUE, auto_unbox = TRUE)

代码说明 #

1. 分组拆分：通过filter拆分数据，避免空值层级干扰嵌套逻辑；
2. 差异化嵌套：
   • 非空组：依次将叶子节点嵌套到Level3、Level2、Level1的children下；
   • 空值组：直接移除Level3列，将叶子节点嵌套到Level2的children下；
3. 合并与转JSON：用bind_rows合并结果，按Level1汇总子节点列表，auto_unbox = TRUE确保不会生成多余的数组包裹。

验证示例结果 #

生成的JSON结构如下，可见空Level3的叶子节点直接挂载至Level2，无空节点冗余：

[
  {
    "Level1": "A",
    "children": [
      {
        "Level2": "A1",
        "children": [
          {
            "Level3": "A1a",
            "children": [
              {
                "Name": "Leaf1",
                "Id": 1
              }
            ]
          }
        ]
      },
      {
        "Level2": "A2",
        "children": [
          {
            "Name": "Leaf2",
            "Id": 2
          }
        ]
      }
    ]
  },
  {
    "Level1": "B",
    "children": [
      {
        "Level2": "B1",
        "children": [
          {
            "Level3": "B1a",
            "children": [
              {
                "Name": "Leaf3",
                "Id": 3
              }
            ]
          }
        ]
      },
      {
        "Level2": "B2",
        "children": [
          {
            "Name": "Leaf4",
            "Id": 4
          }
        ]
      }
    ]
  }
]

内容的提问来源于stack exchange，提问作者Lyla