如何在Ansible中部分扁平化并合并嵌套字典而非完全扁平化？

原始数据 #

我有如下嵌套结构的字典列表：

dict3: [[[{"x": "1","y": "2"},{"z": 3,"h": 4}],{"w": 5}],[[{"x": "333","y": "444"},{"z": "555","h": "777"}],{"w": "999"}]]

目标格式 #

需要将其转换为每个元素是合并后单字典的列表：

dict: [
            {
              "x": "1",
              "y": "2",
              "z": "3",
              "h": "4",
              "w": 5
            },
            {
              "x": "333",
              "y": "444",
              "z": 555,
              "h": 777,
              "w": 999
            }
        ]

当前尝试的问题 #

使用community.general.flattened lookup会将所有元素完全扁平化，但无法合并每组中的x/y/z/h/w字段：

- name: 'unnest' all elements into single list
  ansible.builtin.debug:
    msg: "all in one list {{lookup('community.general.flattened', dict)}}"

扁平化后的结果仅拆分了嵌套结构，未合并字典：

"_raw_params": [
    {
        "x": "1",
        "y": "2"
    },
    {
        "h": 4,
        "z": 3
    },
    {
        "w": 5
    },
    {
        "x": "333",
        "y": "444"
    },
    {
        "h": "777",
        "z": "555"
    },
    {
        "w": "999"
    }
]

当前代码 #

- hosts: localhost
  gather_facts: false
  
  vars:
    dict: [[[{"x": "1","y": "2"},{"z": 3,"h": 4}],{"w": 5}],[[{"x": "333","y": "444"},{"z": "555","h": "777"}],{"w": "999"}]]
  
  tasks:
  - name: UNNEST
    set_fact:
      "{{lookup('community.general.flattened',dict)}}"

解决方案 #

可以通过循环遍历顶层元素，结合flatten和combine过滤器实现字典合并：

- hosts: localhost
  gather_facts: false
  
  vars:
    dict: [[[{"x": "1","y": "2"},{"z": 3,"h": 4}],{"w": 5}],[[{"x": "333","y": "444"},{"z": "555","h": "777"}],{"w": "999"}]]
  
  tasks:
  - name: 合并嵌套字典为目标格式
    set_fact:
      merged_dict: "{{ merged_dict | default([]) + [ item | flatten | combine ] }}"
    loop: "{{ dict }}"

  - name: 查看结果
    debug:
      var: merged_dict

逻辑说明 #

1. loop: "{{ dict }}" 遍历原始列表中的每个顶层嵌套单元（即[[{x,y},{z,h}],{w}]结构）
2. item | flatten 将当前单元的嵌套结构扁平化，得到[{x,y}, {z,h}, {w}]
3. | combine 将扁平化后的多个字典合并为一个完整字典
4. merged_dict | default([]) + [...] 把每个合并后的字典添加到结果列表中

额外说明 #

如果需要对字段类型进行转换（比如目标格式中部分数字字段从字符串转为整数），可以在合并后单独处理，例如：

- name: 转换字段类型（可选）
  set_fact:
    merged_dict: "{{ merged_dict | map('combine', {'z': item.z | int, 'h': item.h | int, 'w': item.w | int}) | list }}"
  loop: "{{ merged_dict }}"
  when: item.z is string or item.h is string or item.w is string

内容的提问来源于stack exchange，提问作者LJS