Project — CSV ↔ JSON Converter

• Build convert.py — a CLI that reads sys.argv for the input and output paths.
• Detect the conversion direction from file extensions (.csv or .json).
• Write the result to disk, preserving column order on the way back to CSV.
• Round-trip: prove that file → other format → original format gives back the same data.

sys.argv is a list of strings — every word you typed after python:

# echo.py
import sys
print(sys.argv)

$ python echo.py hello world
['echo.py', 'hello', 'world']

Index 0 is the script name. The real arguments start at index 1. We'll use sys.argv[1] for the input path and sys.argv[2] for the output.

Before any code, write the recipe:

1. Read inp_path and out_path from sys.argv.
2. Decide direction from extensions.
   - inp .csv  → load CSV into a list of dicts
   - inp .json → load JSON into a list of dicts
3. Pick the output format.
   - out .json → json.dump
   - out .csv  → csv.DictWriter
4. Print a one-line success message.

Step 2 — Load CSV as list of dicts

import csv

def load_csv(path):
    with open(path, newline="") as f:
        return list(csv.DictReader(f))

Step 3 — Save list of dicts as JSON

import json

def save_json(rows, path):
    with open(path, "w") as f:
        json.dump(rows, f, indent=2, ensure_ascii=False)

Step 4 — JSON → CSV

The tricky bit is column order. JSON has no concept of order, but CSV does. Use the keys of the first row, plus a union of every other row's keys (so we don't drop columns that only appear in later rows):

def save_csv(rows, path):
    fields = list(rows[0].keys())
    for r in rows[1:]:
        for k in r:
            if k not in fields:
                fields.append(k)
    with open(path, "w", newline="") as f:
        w = csv.DictWriter(f, fieldnames=fields)
        w.writeheader()
        w.writerows(rows)

Step 5 — The router

Pick which loader and which saver based on extensions:

import sys
from pathlib import Path

inp = Path(sys.argv[1])
out = Path(sys.argv[2])

# Load
if inp.suffix == ".csv":
    rows = load_csv(inp)
elif inp.suffix == ".json":
    with inp.open() as f:
        rows = json.load(f)
else:
    sys.exit(f"❌ unknown input format: {inp.suffix}")

# Save
if out.suffix == ".json":
    save_json(rows, out)
elif out.suffix == ".csv":
    save_csv(rows, out)
else:
    sys.exit(f"❌ unknown output format: {out.suffix}")

print(f"✅ {inp.name}  →  {out.name}  ({len(rows)} rows)")

Assemble everything from the previous section into one file.

# convert.py — CSV ↔ JSON converter
import csv
import json
import sys
from pathlib import Path

def load_csv(path):
    with open(path, newline="") as f:
        return list(csv.DictReader(f))

def load_json(path):
    with open(path) as f:
        return json.load(f)

def save_json(rows, path):
    with open(path, "w") as f:
        json.dump(rows, f, indent=2, ensure_ascii=False)

def save_csv(rows, path):
    fields = list(rows[0].keys())
    for r in rows[1:]:
        for k in r:
            if k not in fields:
                fields.append(k)
    with open(path, "w", newline="") as f:
        w = csv.DictWriter(f, fieldnames=fields)
        w.writeheader()
        w.writerows(rows)

def main(argv):
    if len(argv) != 3:
        sys.exit("usage: python convert.py <in.csv|json> <out.csv|json>")
    inp, out = Path(argv[1]), Path(argv[2])

    if inp.suffix == ".csv":
        rows = load_csv(inp)
    elif inp.suffix == ".json":
        rows = load_json(inp)
    else:
        sys.exit(f"❌ unknown input format: {inp.suffix}")

    if out.suffix == ".json":
        save_json(rows, out)
    elif out.suffix == ".csv":
        save_csv(rows, out)
    else:
        sys.exit(f"❌ unknown output format: {out.suffix}")

    print(f"✅ {inp.name}  →  {out.name}  ({len(rows)} rows)")

if __name__ == "__main__":
    main(sys.argv)

Try it

$ python convert.py students.csv students.json
✅ students.csv  →  students.json  (4 rows)

$ python convert.py students.json roundtrip.csv
✅ students.json  →  roundtrip.csv  (4 rows)

You shipped a real CLI tool. The shape — read args, choose a loader, choose a saver, write output — is the shape of nearly every script in this level. Round-tripping is the test that proves a converter is honest: A → B → A' should equal A. Tomorrow we handle the next universal data problem: dates and times.