Structured Data as Template Input

One value vs. many

Lesson 2's template filled in a hostname and a description — one value each, known in advance. Real device configuration is rarely that simple. A single switch might need three, five, or a dozen VLANs configured, each with its own ID and name. That's not "one more variable" — it's a list of related values, the same shape "Lists & Dictionaries for Network Data" already covered.

device_vlans.py

device = {
  "hostname": "SW1",
  "vlans": [
      {"id": 10, "name": "USERS"},
      {"id": 20, "name": "VOICE"},
      {"id": 30, "name": "MGMT"},
  ],
}

Nothing new here — a dict whose 'vlans' key holds a list of dicts. This is the exact 'list of dictionaries' shape from Python Lesson 3, and the same shape structured inventory has used since Network Automation Fundamentals.

Not a new data concept — a new use for one

This isn't teaching a new way to model data. It's the same list-of-dicts shape you've used since early in this curriculum, now used specifically as a template's input instead of plain program logic.

Quick checkConcept check

What Python data shape best represents a device's multiple VLANs, each with its own ID and name?

A single string listing all the VLAN IDsA list of dictionaries — one dict per VLAN, each with an id and a name keyA separate variable for every possible VLAN IDA dictionary where every key is the word 'vlan'

Flat variables

hostnameone known value
descriptionone known value
Fine for Lesson 2's single interface block

Structured, iterable data

vlans: [ ... ]a list — any number of items
Each item its own dictid, name
What a real device's VLANs actually look like

Accessing nested data inside a template

A template can reach into that structure the same way Python does — dot notation for a dict's keys, and [index] for a list's position:

index_access.py

from jinja2 import Template

device = {
  "hostname": "SW1",
  "vlans": [
      {"id": 10, "name": "USERS"},
      {"id": 20, "name": "VOICE"},
      {"id": 30, "name": "MGMT"},
  ],
}

template = Template("vlan {{ device.vlans[1].id }}\n name {{ device.vlans[1].name }}")
print(template.render(device=device))

device.vlans[1] reaches the second VLAN dict (index 1, the middle one); .id and .name then reach into that dict — dot notation works for dict keys inside a Jinja2 expression, not just object attributes.

This prints:

output

vlan 20
name VOICE

Index 1 is VOICE — the second item in the list (0 = USERS, 1 = VOICE, 2 = MGMT), same zero-based indexing Python itself uses.

Quick checkPredict output

Using the same device dict, what does a template containing vlan {{ device.vlans[2].id }} / name {{ device.vlans[2].name }} render?

vlan 20 / name VOICEvlan 30 / name MGMTvlan 10 / name USERSAn error — index 2 doesn't exist

Quick checkIdentify the correct code

A device dict has a "site" key holding {"name": "HQ-Fl2", "region": "East"}. Which expression correctly reaches the region inside a template?

The awkward part this sets up

Writing device.vlans[0], then device.vlans[1], then device.vlans[2] by hand doesn't scale any better than the hardcoded strings Lesson 1 started with — it just moved the repetition from Python to the template. A device with a dozen VLANs would need a dozen manually-indexed lines. What's actually needed is a way to say "do this once, for every item in the list" — which is exactly what Lesson 4 introduces.

Quick checkScenario

A device has 8 VLANs. Manually indexing device.vlans[0] through device.vlans[7] inside a template works, but what's the real problem with that approach?

Jinja2 has a hard limit of 5 indexed accesses per templateIt doesn't scale — every device with a different number of VLANs needs a hand-edited template, defeating the point of having one reusable templateIndex access doesn't actually work in Jinja2It only works for exactly 3 VLANs

Key takeaways

A device's multiple VLANs are modeled as a list of dicts — the same 'list of dictionaries' shape from Python Lesson 3, now used as template input.
Inside a Jinja2 template, dot notation reaches a dict's keys and [index] reaches a list's items — the same zero-based indexing Python uses.
This is not a new data concept — it's a new use for a data shape this curriculum has used since Network Automation Fundamentals.
Manually indexing every item in a list doesn't scale to devices with a different number of VLANs — the exact problem Lesson 4's loop solves.

What you'll learn

One value vs. many

Accessing nested data inside a template

The awkward part this sets up