jsonl

A lightweight, dependency-free Python library for JSON Lines — read, write, compress, and stream with ease.

Documentation · Changelog · Issues

---

jsonl provides a simple, Pythonic API for working with JSON Lines data. It follows the conventions of Python's standard json module — if you know json.dump and json.load, you already know how to use jsonl.

Fully compliant with the jsonlines and ndjson specifications.

Features

Feature	Description
🌎 Familiar API	Interface similar to the standard json module (dump, load, dumps)
⚡ Streaming by default	Read and write incrementally via iterators, keeping memory usage low
🗜️ Built-in compression	Transparent support for gzip, bzip2, xz, and zst (Python ≥ 3.14)
📦 Archive support	Read and write ZIP and TAR archives (.tar.gz, .tar.bz2, .tar.xz, and .tar.zst (Python ≥ 3.14) )
📥 Load from URLs	Pass a URL directly to load() or load_archive()
🚀 Pluggable serialization	Swap in orjson, or any JSON library
🔧 Error tolerance	Optionally skip malformed lines instead of crashing
🐍 Zero dependencies	Uses only the Python standard library — nothing else

Installation

pip install py-jsonl

Requires Python 3.8+. No external dependencies.

Quick Start

Write

import jsonl

data = [
    {"name": "Gilbert", "wins": [["straight", "7♣"], ["one pair", "10♥"]]},
    {"name": "May", "wins": []},
]

jsonl.dump(data, "players.jsonl")

Read

import jsonl

for item in jsonl.load("players.jsonl"):
    print(item)

Read from a URL

import jsonl

for item in jsonl.load("https://example.com/data.jsonl"):
    print(item)

Compressed files

The compression format is determined automatically — by file extension when writing, and by magic numbers when reading if the file extension is not recognized:

import jsonl

data = [{"key": "value"}]

jsonl.dump(data, "file.jsonl.gz")  # gzip
jsonl.dump(data, "file.jsonl.bz2")  # bzip2
jsonl.dump(data, "file.jsonl.xz")  # xz
jsonl.dump(data, "file.jsonl.zst")  # zst (Python ≥ 3.14) 

for item in jsonl.load("file.jsonl.gz"):
    print(item)

Archives (ZIP / TAR)

import jsonl

# Write multiple files into an archive
data = [
    ("users.jsonl", [{"name": "Alice", "age": 30}, {"name": "Bob", "age": 25}]),
    ("orders.jsonl", [{"id": 1, "total": 99.90}, {"id": 2, "total": 45.00}]),
]
jsonl.dump_archive("data.tar.gz", data)

# Read them back
for filename, items in jsonl.load_archive("data.tar.gz"):
    print(f"--- {filename} ---")
    for item in items:
        print(item)

Multiple output files

import jsonl

data = [
    ("file1.jsonl", [{"name": "Alice"}, {"name": "Bob"}]),
    ("file2.jsonl", [{"name": "Charlie"}]),
    ("file1.jsonl", [{"name": "Eve"}]),  # appended to file1.jsonl
]

jsonl.dump_fork(data)

API Overview

Reading

Function	Description
jsonl.load(source, **kw)	Read from a file, URL, or file-like object
jsonl.load_archive(file, **kw)	Unpack JSON Lines files from a ZIP or TAR archive
jsonl.loader(stream, broken, **kw)	Low-level generator deserializing a line stream

Tip

All read functions accept cls and **kwargs for custom decoding.

Writing

Function	Description
jsonl.dump(iterable, file, **kw)	Write objects to a JSON Lines file
jsonl.dumps(iterable, **kw)	Serialize to a JSON Lines string
jsonl.dump_fork(paths, **kw)	Write to multiple JSON Lines files at once
jsonl.dump_archive(path, data, **kw)	Pack multiple JSON Lines files into a ZIP or TAR archive
jsonl.dumper(iterable, **kw)	Low-level generator yielding formatted lines

Tip

All write functions accept cls and **kwargs for custom encoding.

For complete parameter documentation, see the full docs →

Custom Serialization

Plug in any JSON-compatible serializer.

For example, orjson for high-performance encoding:

import orjson  # ensure orjson is installed: pip install orjson
import jsonl

data = [{"name": "Alice", "age": 30}, {"name": "Bob", "age": 25}]

# Write with orjson (returns bytes → set text_mode=False)
jsonl.dump(data, "fast.jsonl", text_mode=False, cls=orjson.dumps)

# Read with orjson
for item in jsonl.load("fast.jsonl", cls=orjson.loads):
    print(item)

Another example: using custom cls with **kwargs for various purposes, for example:

import datetime
import decimal
import json

import jsonl


class UpperDecoder(json.JSONDecoder):
    def __init__(self, *args, **kwargs):
        super().__init__(*args, object_hook=self.object_hook, parse_float=decimal.Decimal, **kwargs)

    def object_hook(self, obj):
        return {k.upper(): v for k, v in obj.items()}


class ISODateEncoder(json.JSONEncoder):
    def default(self, obj):
        if isinstance(obj, datetime.date):
            return obj.isoformat()
        return super().default(obj)


data = [
    {"name": "Alice", "birthdate": datetime.date(2000, 1, 1)},
    {"name": "Bob", "birthdate": datetime.date(2005, 1, 1)}
]

#  Write using a custom encoder to serialize datetime objects as ISO strings
jsonl.dump(data, "file.jsonl", cls=ISODateEncoder)

# Read using a custom decoder to convert floats into Decimal and uppercase all keys
for item in jsonl.load("file.jsonl", cls=UpperDecoder):
    print(item)

keyword arguments are forwarded to the underlying serializer:

import jsonl

data = [{"name": "Alice", "score": 9.5}, {"name": "Bob", "score": 7.2}]

jsonl.dump(data, "compact.jsonl", separators=(",", ":"))  # compact output
jsonl.dump(data, "sorted.jsonl", sort_keys=True)  # deterministic keys

Supported Formats

Type	Extensions
Plain	.jsonl
Compressed	.jsonl.gz, .jsonl.bz2, .jsonl.xz, .jsonl.zst (Python ≥ 3.14)
ZIP archive	.zip
TAR archive	.tar, .tar.gz, .tar.bz2, .tar.xz, .tar.zst (Python ≥ 3.14)

When reading, if the file extension is not recognized, jsonl falls back to magic-number detection to identify the compression format automatically.

Contributing

# Install dev dependencies
pip install --group=test --upgrade

# Run tests
python -Wd -m pytest tests/
python -Wd -m pytest tests/ --cov  # run with coverage reporting

# Lint
pip install --group=lint --upgrade
ruff check .

# Docs
pip install --group=doc --upgrade

# zensical usage: https://zensical.org/docs/usage/
zensical build 
zensical serve

License

MIT — see LICENSE for details.