JSON — Nested Data & Pretty Output

• Read deeply nested JSON: dict-of-lists-of-dicts.
• Walk a JSON tree to extract specific fields.
• Use indent, sort_keys and ensure_ascii to control output.
• Update one nested value and write the file back.

This is the shape every dataset eventually takes:

{
  "school": "SMK Sungai Long",
  "year":   2026,
  "classes": [
    {
      "name": "Form 1A",
      "students": [
        {"id": 1, "name": "Aisyah",  "scores": [88, 92, 79]},
        {"id": 2, "name": "Wei Jie", "scores": [70, 75, 81]}
      ]
    },
    {
      "name": "Form 1B",
      "students": [
        {"id": 3, "name": "Suresh",  "scores": [95, 90, 88]}
      ]
    }
  ]
}

Predict: how would you reach Wei Jie's first score (70)?

data["classes"][0]["students"][1]["scores"][0]

Save the data file

Create school.json with the JSON from the warm-up. Then read it:

import json
with open("school.json") as f:
    school = json.load(f)

Walk it with nested for loops

for cls in school["classes"]:
    print(f"\nClass: {cls['name']}")
    for student in cls["students"]:
        avg = sum(student["scores"]) / len(student["scores"])
        print(f"  {student['name']:<10} avg = {avg:.1f}")

Class: Form 1A
  Aisyah     avg = 86.3
  Wei Jie    avg = 75.3

Class: Form 1B
  Suresh     avg = 91.0

Pretty-print kwargs

json.dump and json.dumps take three formatting kwargs:

indent=2          → readable indentation
sort_keys=True    → keys in alphabetical order
ensure_ascii=False → keep Unicode (中, é, 🚀) as-is

Compare:

import json
d = {"name": "Café", "type": "Kopitiam", "rating": 4.5}

print(json.dumps(d))
print("---")
print(json.dumps(d, indent=2))
print("---")
print(json.dumps(d, indent=2, sort_keys=True, ensure_ascii=False))

{"name": "Caf\u00e9", "type": "Kopitiam", "rating": 4.5}
---
{
  "name": "Caf\u00e9",
  "type": "Kopitiam",
  "rating": 4.5
}
---
{
  "name": "Café",
  "rating": 4.5,
  "type": "Kopitiam"
}

The default escapes non-ASCII to \\uXXXX. Set ensure_ascii=False when you want real characters in the file.

Modify and write back

Reading + editing + saving is a classic three-step pattern:

with open("school.json") as f:
    school = json.load(f)

# Edit: add a new score to Aisyah
school["classes"][0]["students"][0]["scores"].append(95)

with open("school.json", "w") as f:
    json.dump(school, f, indent=2, ensure_ascii=False)

Build a one-page printable report from the school JSON.

# class_report.py — pretty report from nested JSON
import json

with open("school.json") as f:
    school = json.load(f)

print(f"📘 {school['school']} — {school['year']}")
print("=" * 40)

for cls in school["classes"]:
    print(f"\n🏫 {cls['name']}  ({len(cls['students'])} students)")
    print("-" * 40)
    for student in cls["students"]:
        scores = student["scores"]
        avg = sum(scores) / len(scores)
        best = max(scores)
        print(f"  #{student['id']:<2} {student['name']:<10} "
              f"avg={avg:5.1f}  best={best}")

# Save a summary section back into the JSON
school["summary"] = {
    "total_students": sum(len(c["students"]) for c in school["classes"]),
    "class_count":    len(school["classes"]),
}
with open("school.json", "w") as f:
    json.dump(school, f, indent=2, ensure_ascii=False)

print(f"\n📝 summary block written.")

Sample output

📘 SMK Sungai Long — 2026
========================================

🏫 Form 1A  (2 students)
----------------------------------------
  #1  Aisyah     avg= 86.3  best=92
  #2  Wei Jie    avg= 75.3  best=81

🏫 Form 1B  (1 students)
----------------------------------------
  #3  Suresh     avg= 91.0  best=95

📝 summary block written.

Read the diff

Two patterns to notice. The nested loop walks classes-then-students. The aggregating comprehension sum(len(c["students"]) for c in school["classes"]) totals across the whole tree in one expression. That's how data work feels in Python.

Nested JSON is read by chaining [key] and [index] until you reach the value. To process the whole tree, nest for loops one per level. The formatting kwargs (indent, sort_keys, ensure_ascii) control how the output looks on disk. Read + modify + write is the standard pattern.