json

Standard library module View source

Native JSON parsing and encoding with a tagged JSON value.

Provides load, loads, dump, dumps, JSONDecoder, JSONEncoder, and JSONDecodeError. Standard formatting options and non-finite number defaults follow Python.

import json

value = json.loads('{"items": [1, true, null]}')
number = value["items"][0].as_int()
value["items"].append("text")
print(json.dumps(value, indent=2))

Differences from Python:

• Decoding returns JSON. Use kind, is_null(), and typed accessors as_bool(), as_int(), as_float(), as_str(); wrong kinds raise TypeError (as_float() also accepts integers). JSON() is null; JSON(value) converts supported native scalars, optionals, lists, tuples, and string-keyed dicts. JSON.array() and JSON.object() create empty containers.
• Use values(), keys(), or items() instead of direct iteration. Containers support indexing and mutation; copies share storage. Scalars are immutable. Equality is structural and kind-sensitive (JSON(1) != JSON(1.0)); values are unhashable. str(value) serializes JSON, including quotes around strings.
• Integers are signed 64-bit; oversized literals raise OverflowError. Fractional and exponent-form numbers use binary64.
• Keys must be strings. skipkeys, cls, default, object_hook, object_pairs_hook, parse_int, parse_float, and parse_constant are unsupported; unsupported native types, including nested members, are compile-time errors.
• Parsing/encoding use max_depth=512, raising RecursionError when exceeded, even with check_circular=False. Structural comparison also has a 512 limit.
• I/O is not streaming: load() reads the whole stream; dump() writes a complete string to a text stream. iterencode() is not provided.
• Decode-error locations use Unicode code points; trailing-comma diagnostics follow CPython 3.14.

class JSON

@tuple

class JSON

View complete class API →

Two-word value with a packed tag and string metadata, plus scalar bits or a pointer.

Constructors 2

__new__

__new__()

__new__

__new__(value)

Properties 1

kind

@property

kind(self)

Methods 17

array

array()

object

object()

is_null

is_null(self)

is_number

is_number(self)

is_bool

is_bool(self)

is_str

is_str(self)

is_array

is_array(self)

is_object

is_object(self)

as_bool

as_bool(self)

as_int

as_int(self)

as_float

as_float(self)

as_str

as_str(self)

append

append(self, value)

get

get(self, key: str, default = None)

values

values(self)

keys

keys(self)

items

items(self)

Special methods 13

__int__

__int__(self)

__float__

__float__(self)

__bool__

__bool__(self)

__len__

__len__(self)

__getitem__

__getitem__(self, index: int)

__getitem__(self, key: str)

__setitem__

__setitem__(self, index: int, value)

__setitem__(self, key: str, value)

__delitem__

__delitem__(self, index: int)

__delitem__(self, key: str)

__contains__

__contains__(self, key: str)

__str__

__str__(self)

__repr__

__repr__(self)

__eq__

__eq__(self, other: JSON)

__ne__

__ne__(self, other: JSON)

__hash__

__hash__(self)

class JSONDecodeError

class JSONDecodeError

View complete class API →

Fields 5

msg str

doc str

pos int

lineno int

colno int

Constructors 1

__init__

__init__(self, msg: str, doc: str, pos: int)

class JSONDecoder

class JSONDecoder

View complete class API →

Fields 2

strict bool

max_depth int

Constructors 1

__init__

__init__(self, strict: bool = True, max_depth: int = 512)

Methods 2

raw_decode

raw_decode(self, source: str, idx: int = 0)

decode

decode(self, source: str)

function loads

loads(source, strict: bool = True, max_depth: int = 512)

function load

load(fp, strict: bool = True, max_depth: int = 512)

class JSONEncoder

class JSONEncoder

View complete class API →

Fields 8

ensure_ascii bool

allow_nan bool

check_circular bool

sort_keys bool

pretty bool

indent str

separators Tuple[str, str]

max_depth int

Constructors 1

__init__

__init__(self, ensure_ascii: bool = True, check_circular: bool = True, allow_nan: bool = True, indent = None, separators: Optional[Tuple[str, str]] = None, sort_keys: bool = False, max_depth: int = 512)

Methods 1

encode

encode(self, value)

function dumps

dumps(value, ensure_ascii: bool = True, check_circular: bool = True, allow_nan: bool = True, indent = None, separators: Optional[Tuple[str, str]] = None, sort_keys: bool = False, max_depth: int = 512)

function dump

dump(value, fp, ensure_ascii: bool = True, check_circular: bool = True, allow_nan: bool = True, indent = None, separators: Optional[Tuple[str, str]] = None, sort_keys: bool = False, max_depth: int = 512)