Dynamically generating complex data in Ansible - list

Question

I'm facing the following problem set in writing an Ansible playbook:
A list is given to me:
rooms:
- room_name: bedroom-1
chairs: 1
- room_name: bedroom-2
chairs: 0
- room_name: bathroom
chairs: 0
- room_name: kitchen
chairs: 4
And I need to transform that into a list of dictionaries with these two keys per item:
chair_name: <room_name>-chair-<chair_number>
room_name: <room_name>
With the example above that would give me:
chair_names:
- chair_name: bedroom-1-chair-1
room_name: bedroom-1
- chair_name: kitchen-chair-1
room_name: kitchen
- chair_name: kitchen-chair-2
room_name: kitchen
- chair_name: kitchen-chair-3
room_name: kitchen
- chair_name: kitchen-chair-4
room_name: kitchen
I've been struggling to accomplish this using Ansible. Is there a way to do it?

Answer 1

Use Jinja to create the structure
chair_names_str: |-
{% for room in rooms %}
{% for i in range(1, room.chairs + 1) %}
- chair_name: {{ room.room_name }}-chair-{{ i }}
room_name: {{ room.room_name }}
{% endfor %}
{% endfor %}
chair_names: "{{ chair_names_str|from_yaml }}"
gives
chair_names:
- {chair_name: bedroom-1-chair-1, room_name: bedroom-1}
- {chair_name: kitchen-chair-1, room_name: kitchen}
- {chair_name: kitchen-chair-2, room_name: kitchen}
- {chair_name: kitchen-chair-3, room_name: kitchen}
- {chair_name: kitchen-chair-4, room_name: kitchen}
Q: "Is it possible to achieve the same results using Ansible loops?"
A: Yes. It is. In the first loop create lists of the chairs
- set_fact:
chairs: "{{ chairs|d([]) + [{'chairs': range(1, item|int + 1)|list}] }}"
loop: "{{ rooms|map(attribute='chairs')|list }}"
gives
chairs:
- chairs: [1]
- chairs: []
- chairs: []
- chairs: [1, 2, 3, 4]
In the second loop create the list. The task below gives the same result
- set_fact:
chair_names: "{{ chair_names|d([]) + [_item] }}"
with_subelements:
- "{{ rooms|zip(chairs)|map('combine') }}"
- chairs
vars:
_item:
chair_name: "{{ item.0.room_name }}-chair-{{ item.1 }}"
room_name: "{{ item.0.room_name }}"
You can see that the first option is much simpler.