json-lines 0.5.0

Why?

Reading a well-formed JSON lines file is a one-liner in Python. But the file can be broken: cut at some point (this happens when the process writing it is killed), or concatenated from several cut pieces (this happend when the process started appending to the same file again). Handling all this cases is not easy, especially if the file is compressed.

json-lines handles all this cases for you!

Installation

pip install json-lines

If ujson is installed, it is used to speed up json decoding (which is the main performance bottleneck even for gzipped files).

Usage

In order to read a well-formed json lined file, pass an open file as the first argument to json_lines.reader. The file can be opened in text or binary mode, but if it’s opened in text mode, the encoding must be set correctly:

import json_lines

with open('file.jl', 'rb') as f:
    for item in json_lines.reader(f):
        print(item['x'])

There is also a helper function json_lines.open that recognizes “.gz” and “.gzip” extensions and opens them with gzip:

with json_lines.open('file.jl.gz') as f:
    for item in f:
        print(item['x'])

Handling broken (cut at some point) files is enabled by passing broken=True to json_lines.reader or json_lines.open. Broken lines are skipped (only logging a warning), and reading continues from the next valid position. This works both for compressed and uncompressed files:

with json_lines.open('file.jl.gz', broken=True) as f:
    for item in f:
        print(item['x'])

License

License is MIT.