Ansible Nested Lists: Loop Over Lists of Lists (with_subelements, flatten)

Introduction

CfgMgmtCamp 2025, the premier event for systems administrators, DevOps engineers, and Infrastructure as Code enthusiasts, is taking place in Ghent, Belgium, from February 3–5, 2025. Among the many notable speakers, Luca Berton, a globally recognized Ansible expert, will share his expertise on automating AI-driven environments.

See also: Luca Berton at CfgMgmtCamp 2025: Ansible Automation Expert on Neo4j GenAI

Session Overview: Automating AI-Powered Graph Databases with Ansible

Title: Automating AI-Powered Graph Databases with Ansible: A Neo4j GenAI Case Study Date & Time: February 4, 2025, 14:00–14:25 Location: Ansible 1 (B.1.017) Session Type: Short Talk (25 minutes)

In this insightful session, Luca Berton will demonstrate: • Automating the deployment of Neo4j GenAI environments using Ansible playbooks. • Integrating OpenAI for retrieval-augmented generation (RAG) tasks. • Best practices for configuring, managing, and optimizing AI-powered graph databases in hybrid cloud environments.

This talk is tailored for IT professionals looking to streamline complex AI setups using modern Infrastructure as Code (IaC) techniques. Expect hands-on demonstrations and actionable takeaways to enhance your automation journey.

About Luca Berton

Luca Berton is a seasoned Ansible Automation Expert and the author of best-selling books, including: • Ansible for Kubernetes by Example: Automate Your Kubernetes Cluster • Ansible for VMware by Example

With over 15 years of experience, Luca has been a key contributor to the open-source community. He has worked with Red Hat and JPMorgan Chase & Co., leading projects in automation, infrastructure hardening, and hybrid cloud deployments.

See also: Automating AI-Powered Graph Databases with Ansible: Insights from CfgMgmtCamp 2025

Don’t Miss This Event

CfgMgmtCamp 2025 offers a unique platform to connect with thought leaders and explore the latest in configuration management, automation, and DevOps. Luca Berton's session promises to be a highlight for anyone passionate about cutting-edge AI and Ansible automation.

For more details and registration, visit CfgMgmtCamp 2025.

---

Resources

Check out Luca Berton’s books and resources to dive deeper into Ansible automation: • Ansible by Example: 200+ Automation Examples • Ansible for Kubernetes by Example

Want to support Luca’s work? Donate on Patreon.

See also: Setting Up Neo4j GenAI Environment on Fedora Using Ansible

Working with Nested Lists

Flatten nested lists

- name: Flatten nested list
  vars:
    nested_packages:
      - [nginx, curl]
      - [git, vim]
      - [htop]
  ansible.builtin.apt:
    name: "{{ nested_packages | flatten }}"
    state: present
  become: true
  # Installs: nginx, curl, git, vim, htop

Loop over list of dictionaries with sub-lists

- name: Create users with multiple groups
  vars:
    users:
      - name: alice
        groups: [developers, docker]
      - name: bob
        groups: [developers, sudo]
      - name: charlie
        groups: [ops, docker, sudo]
  ansible.builtin.user:
    name: "{{ item.name }}"
    groups: "{{ item.groups | join(',') }}"
    append: true
  loop: "{{ users }}"
  become: true

with_subelements (nested iteration)

- name: Add SSH keys for each user
  vars:
    users:
      - name: alice
        keys:
          - "ssh-rsa AAAA... alice@laptop"
          - "ssh-rsa BBBB... alice@desktop"
      - name: bob
        keys:
          - "ssh-rsa CCCC... bob@work"
  ansible.posix.authorized_key:
    user: "{{ item.0.name }}"
    key: "{{ item.1 }}"
  with_subelements:
    - "{{ users }}"
    - keys
  become: true

Nested loop (cartesian product)

- name: Create directories for each app in each environment
  vars:
    apps: [frontend, backend, api]
    envs: [dev, staging, prod]
  ansible.builtin.file:
    path: "/opt/{{ item.1 }}/{{ item.0 }}"
    state: directory
    mode: '0755'
  with_nested:
    - "{{ apps }}"
    - "{{ envs }}"
  become: true
  # Creates: /opt/dev/frontend, /opt/dev/backend, ..., /opt/prod/api

Modern loop with product filter

- name: Same as above using loop + product
  ansible.builtin.file:
    path: "/opt/{{ item.1 }}/{{ item.0 }}"
    state: directory
    mode: '0755'
  loop: "{{ apps | product(envs) | list }}"
  become: true

Accessing Nested Data

vars:
  servers:
    web:
      - host: web1.example.com
        port: 80
      - host: web2.example.com
        port: 8080
    db:
      - host: db1.example.com
        port: 5432
# Access nested values
- debug:
    msg: "{{ servers.web[0].host }}:{{ servers.web[0].port }}"
  # Output: web1.example.com:80
# Loop over dict + sub-list
- debug:
    msg: "{{ item.key }}: {{ item.value | map(attribute='host') | join(', ') }}"
  loop: "{{ servers | dict2items }}"

FAQ

When should I use with_subelements vs loop + subelements filter?

They're equivalent. The filter syntax is more modern:

# Old style
with_subelements:
  - "{{ users }}"
  - keys
# Modern style
loop: "{{ users | subelements('keys') }}"

How do I flatten only one level deep?

# flatten(1) only removes one nesting level
- debug:
    msg: "{{ [[1, [2, 3]], [4, 5]] | flatten(1) }}"
  # Output: [1, [2, 3], 4, 5]

How do I handle optional sub-lists?

Use default([]) to handle missing keys:

loop: "{{ users | subelements('keys', skip_missing=True) }}"

Related Articles

• Ansible flatten Filter: Flatten Nested Lists & Remove Duplicates