How can I read and write a JSON Lines file one record per line?

A JSON Lines file is a sequence of separate lines, with each line holding one complete JSON value. For example, a file might contain:

{"name": "Mina", "score": 8}
{"name": "Omar", "score": 6}

That is different from a single JSON array such as [{...}, {...}]. Python’s json module handles JSON documents, but repeatedly dumping separate values into one file does not make an ordinary valid JSON document. For line-oriented records, read and write one value at a time. [S1]

The example below assumes one complete value per physical line. It skips blank lines as an explicit convenience; if blank lines matter to your workflow, choose and document a different policy.

Use a line counter and catch json.JSONDecodeError for each line. This example keeps valid records and reports errors to the screen rather than stopping at the first bad record:

import json

records = []

with open("input.jsonl", "r", encoding="utf-8") as infile:
    for line_number, line in enumerate(infile, start=1):
        if not line.strip():
            continue  # Explicit policy: ignore blank lines.

        try:
            record = json.loads(line)
        except json.JSONDecodeError as exc:
            print(
                f"Malformed JSON on line {line_number}: "
                f"{exc.msg} (column {exc.colno})"
            )
            continue

        records.append(record)

print(f"Read {len(records)} valid records")

The counter identifies the physical line in the file, while the exception provides parsing details such as its message and position. [S1] If the file is large, you can process each record inside the loop instead of retaining every value in records.

Open the destination in write mode and dump each value separately, followed by a newline:

import json

records = [
    {"name": "Mina", "score": 8},
    {"name": "Omar", "score": 6},
]

with open("output.jsonl", "w", encoding="utf-8") as outfile:
    for record in records:
        outfile.write(json.dumps(record))
        outfile.write("\n")

The result has one serialized value per line. This example replaces the destination file because it uses "w"; use a separate output path if you need to preserve the input. The standard JSON module supplies the serialization, while the added newline supplies the record boundary for this line-oriented approach. [S1]

Yes. Read each input line, parse it, and write valid values to a different output file. Keeping the paths separate avoids replacing the source while it is being read:

import json

with open("input.jsonl", "r", encoding="utf-8") as infile, \
     open("cleaned.jsonl", "w", encoding="utf-8") as outfile:
    for line_number, line in enumerate(infile, start=1):
        if not line.strip():
            continue  # This example ignores blank lines.

        try:
            value = json.loads(line)
        except json.JSONDecodeError as exc:
            print(f"Line {line_number}: {exc.msg} (column {exc.colno})")
            continue

        outfile.write(json.dumps(value) + "\n")

This copies valid JSON values into a new line-oriented file and reports skipped malformed lines. It does not establish requirements for every JSON Lines implementation, nor does it verify that values have the fields or types your application expects. Add application-specific validation if needed, and decide whether skipping blank or malformed lines is appropriate before relying on the output.