How can I sort a JSON list of records by a chosen field?

In Python, a JSON array becomes a list, and each JSON object becomes a dictionary. You can sort that list with a key function, then write the result to another file with json.dump. [S1]

Here is a complete example. It creates a small input file, sorts by numeric score in ascending order, and writes records-sorted.json separately:

import json

records = [
    {"name": "Mina", "score": 12},
    {"name": "Omar", "score": 5},
    {"name": "Lee", "score": 9},
]

with open("records.json", "w", encoding="utf-8") as f:
    json.dump(records, f, indent=2)

with open("records.json", encoding="utf-8") as f:
    data = json.load(f)

field = "score"

if not isinstance(data, list):
    raise ValueError("Expected a JSON array")

for index, record in enumerate(data):
    if not isinstance(record, dict):
        raise ValueError(f"Record {index} is not an object")
    if field not in record:
        raise ValueError(f"Record {index} is missing {field!r}")
    value = record[field]
    if isinstance(value, bool) or not isinstance(value, (int, float)):
        raise ValueError(f"Record {index} has a nonnumeric {field!r}")

sorted_records = sorted(data, key=lambda record: record[field])

with open("records-sorted.json", "w", encoding="utf-8") as f:
    json.dump(sorted_records, f, indent=2)

with open("records-sorted.json", encoding="utf-8") as f:
    saved = json.load(f)

assert [record[field] for record in saved] == [5, 9, 12]
print("Verified sorted output")

The validation makes the policy explicit: a missing field or a value that is not a number stops the run instead of silently choosing an order.

Choose text ordering deliberately, and validate that every record has a string for the field. For example, replace the numeric check and sort line with:

if not isinstance(value, str):
    raise ValueError(f"Record {index} has a non-text {field!r}")

sorted_records = sorted(data, key=lambda record: record[field])

This makes the field type explicit. If you need particular rules for case, accents, or language-specific ordering, define those rules rather than assuming the default matches your needs.

Either policy can be reasonable, but choose it explicitly. Stopping with an error is useful when every record is supposed to have the field, because it exposes malformed input. If missing values are expected, one simple policy is to separate records that have the field from those that do not, sort the first group, and append the missing group. Validate the present values’ types before sorting; do not mix missing values, numbers, and text in one comparison key and rely on accidental ordering. [S1]

Reload the output file and inspect the chosen field in sequence, as the example’s assertion does. That checks the saved file rather than only the in-memory list. Also, sort_keys=True in json.dump sorts dictionary keys when serializing; it does not reorder the list of records. [S1] For a different expected order, change the assertion to match your chosen field and direction. For descending numeric order, use reverse=True in the sorted call and update the verification sequence.