18.2. json — JSON 编码和解码器

2.6 新版功能.

JSON (JavaScript Object Notation),由 RFC 7159 (which obsoletes RFC 4627) 和 ECMA-404 指定,是一个受 JavaScript 的对象字面量语法启发的轻量级数据交换格式,尽管它不仅仅是一个严格意义上的 JavaScript 的字集 [1]

json 提供了与标准库 marshalpickle 相似的API接口。

对基本的 Python 对象层次结构进行编码:

>>> import json
>>> json.dumps(['foo', {'bar': ('baz', None, 1.0, 2)}])
'["foo", {"bar": ["baz", null, 1.0, 2]}]'
>>> print json.dumps("\"foo\bar")
"\"foo\bar"
>>> print json.dumps(u'\u1234')
"\u1234"
>>> print json.dumps('\\')
"\\"
>>> print json.dumps({"c": 0, "b": 0, "a": 0}, sort_keys=True)
{"a": 0, "b": 0, "c": 0}
>>> from StringIO import StringIO
>>> io = StringIO()
>>> json.dump(['streaming API'], io)
>>> io.getvalue()
'["streaming API"]'

紧凑编码:

>>> import json
>>> json.dumps([1,2,3,{'4': 5, '6': 7}], separators=(',',':'))
'[1,2,3,{"4":5,"6":7}]'

美化输出:

>>> import json
>>> print json.dumps({'4': 5, '6': 7}, sort_keys=True,
...                  indent=4, separators=(',', ': '))
{
    "4": 5,
    "6": 7
}

JSON解码:

>>> import json
>>> json.loads('["foo", {"bar":["baz", null, 1.0, 2]}]')
[u'foo', {u'bar': [u'baz', None, 1.0, 2]}]
>>> json.loads('"\\"foo\\bar"')
u'"foo\x08ar'
>>> from StringIO import StringIO
>>> io = StringIO('["streaming API"]')
>>> json.load(io)
[u'streaming API']

特殊JSON对象解码:

>>> import json
>>> def as_complex(dct):
...     if '__complex__' in dct:
...         return complex(dct['real'], dct['imag'])
...     return dct
...
>>> json.loads('{"__complex__": true, "real": 1, "imag": 2}',
...     object_hook=as_complex)
(1+2j)
>>> import decimal
>>> json.loads('1.1', parse_float=decimal.Decimal)
Decimal('1.1')

扩展 JSONEncoder:

>>> import json
>>> class ComplexEncoder(json.JSONEncoder):
...     def default(self, obj):
...         if isinstance(obj, complex):
...             return [obj.real, obj.imag]
...         # Let the base class default method raise the TypeError
...         return json.JSONEncoder.default(self, obj)
...
>>> json.dumps(2 + 1j, cls=ComplexEncoder)
'[2.0, 1.0]'
>>> ComplexEncoder().encode(2 + 1j)
'[2.0, 1.0]'
>>> list(ComplexEncoder().iterencode(2 + 1j))
['[', '2.0', ', ', '1.0', ']']

Using json.tool from the shell to validate and pretty-print:

$ echo '{"json":"obj"}' | python -m json.tool
{
    "json": "obj"
}
$ echo '{1.2:3.4}' | python -mjson.tool
Expecting property name enclosed in double quotes: line 1 column 2 (char 1)

注解

JSON 是 YAML 1.2 的一个子集。由该模块的默认设置生成的 JSON (尤其是默认的 “分隔符” 设置值)也是 YAML 1.0 and 1.1 的一个子集。因此该模块也能够用于序列化为 YAML。

18.2.1. 基本使用

json.dump(obj, fp, skipkeys=False, ensure_ascii=True, check_circular=True, allow_nan=True, cls=None, indent=None, separators=None, encoding="utf-8", default=None, sort_keys=False, **kw)

使用这个 conversion table 来序列化 obj 为一个 JSON 格式的流并输出到 fp (一个支持 .write()file-like object)。

If skipkeys is true (default: False), then dict keys that are not of a basic type (str, unicode, int, long, float, bool, None) will be skipped instead of raising a TypeError.

If ensure_ascii is true (the default), all non-ASCII characters in the output are escaped with \uXXXX sequences, and the result is a str instance consisting of ASCII characters only. If ensure_ascii is false, some chunks written to fp may be unicode instances. This usually happens because the input contains unicode strings or the encoding parameter is used. Unless fp.write() explicitly understands unicode (as in codecs.getwriter()) this is likely to cause an error.

如果 check_circular 是为假值 (默认为 True),那么容器类型的循环引用检验会被跳过并且循环引用会引发一个 OverflowError (或者更糟的情况)。

如果 allow_nan 是 false(默认为 True),那么在对严格 JSON 规格范围外的 float 类型值(naninf-inf)进行序列化时会引发一个 ValueError。如果 allow_nan 是 true,则使用它们的 JavaScript 等价形式(NaNInfinity-Infinity)。

If indent is a non-negative integer, then JSON array elements and object members will be pretty-printed with that indent level. An indent level of 0, or negative, will only insert newlines. None (the default) selects the most compact representation.

注解

Since the default item separator is ', ', the output might include trailing whitespace when indent is specified. You can use separators=(',', ': ') to avoid this.

If specified, separators should be an (item_separator, key_separator) tuple. By default, (', ', ': ') are used. To get the most compact JSON representation, you should specify (',', ':') to eliminate whitespace.

encoding is the character encoding for str instances, default is UTF-8.

default 被指定时,其应该是一个函数,每当某个对象无法被序列化时它会被调用。它应该返回该对象的一个可以被 JSON 编码的版本或者引发一个 TypeError。如果没有被指定,则会直接引发 TypeError

如果 sort_keys 是 true(默认为 False),那么字典的输出会以键的顺序排序。

为了使用一个自定义的 JSONEncoder 子类(比如:覆盖了 default() 方法来序列化额外的类型), 通过 cls 关键字参数来指定;否则将使用 JSONEncoder

注解

Unlike pickle and marshal, JSON is not a framed protocol so trying to serialize more objects with repeated calls to dump() and the same fp will result in an invalid JSON file.

json.dumps(obj, skipkeys=False, ensure_ascii=True, check_circular=True, allow_nan=True, cls=None, indent=None, separators=None, encoding="utf-8", default=None, sort_keys=False, **kw)

Serialize obj to a JSON formatted str using this conversion table. If ensure_ascii is false, the result may contain non-ASCII characters and the return value may be a unicode instance.

The arguments have the same meaning as in dump().

注解

JSON 中的键-值对中的键永远是 str 类型的。当一个对象被转化为 JSON 时,字典中所有的键都会被强制转换为字符串。这所造成的结果是字典被转换为 JSON 然后转换回字典时可能和原来的不相等。换句话说,如果 x 具有非字符串的键,则有 loads(dumps(x)) != x

json.load(fp[, encoding[, cls[, object_hook[, parse_float[, parse_int[, parse_constant[, object_pairs_hook[, **kw]]]]]]]])

Deserialize fp (a .read()-supporting file-like object containing a JSON document) to a Python object using this conversion table.

If the contents of fp are encoded with an ASCII based encoding other than UTF-8 (e.g. latin-1), then an appropriate encoding name must be specified. Encodings that are not ASCII based (such as UCS-2) are not allowed, and should be wrapped with codecs.getreader(encoding)(fp), or simply decoded to a unicode object and passed to loads().

object_hook 是一个可选的函数,它会被调用于每一个解码出的对象字面量(即一个 dict)。object_hook 的返回值会取代原本的 dict。这一特性能够被用于实现自定义解码器(如 JSON-RPC 的类型提示)。

object_pairs_hook is an optional function that will be called with the result of any object literal decoded with an ordered list of pairs. The return value of object_pairs_hook will be used instead of the dict. This feature can be used to implement custom decoders that rely on the order that the key and value pairs are decoded (for example, collections.OrderedDict() will remember the order of insertion). If object_hook is also defined, the object_pairs_hook takes priority.

在 2.7 版更改: Added support for object_pairs_hook.

parse_float, if specified, will be called with the string of every JSON float to be decoded. By default, this is equivalent to float(num_str). This can be used to use another datatype or parser for JSON floats (e.g. decimal.Decimal).

parse_int, if specified, will be called with the string of every JSON int to be decoded. By default, this is equivalent to int(num_str). This can be used to use another datatype or parser for JSON integers (e.g. float).

parse_constant, if specified, will be called with one of the following strings: '-Infinity', 'Infinity', 'NaN'. This can be used to raise an exception if invalid JSON numbers are encountered.

在 2.7 版更改: parse_constant doesn’t get called on ‘null’, ‘true’, ‘false’ anymore.

To use a custom JSONDecoder subclass, specify it with the cls kwarg; otherwise JSONDecoder is used. Additional keyword arguments will be passed to the constructor of the class.

json.loads(s[, encoding[, cls[, object_hook[, parse_float[, parse_int[, parse_constant[, object_pairs_hook[, **kw]]]]]]]])

Deserialize s (a str or unicode instance containing a JSON document) to a Python object using this conversion table.

If s is a str instance and is encoded with an ASCII based encoding other than UTF-8 (e.g. latin-1), then an appropriate encoding name must be specified. Encodings that are not ASCII based (such as UCS-2) are not allowed and should be decoded to unicode first.

The other arguments have the same meaning as in load().

18.2.2. Encoders and Decoders

class json.JSONDecoder([encoding[, object_hook[, parse_float[, parse_int[, parse_constant[, strict[, object_pairs_hook]]]]]]])

Simple JSON decoder.

Performs the following translations in decoding by default:

JSON Python
object dict
array list
string unicode
整数 int, long
非整数 float
true True
false False
null None

It also understands NaN, Infinity, and -Infinity as their corresponding float values, which is outside the JSON spec.

encoding determines the encoding used to interpret any str objects decoded by this instance (UTF-8 by default). It has no effect when decoding unicode objects.

Note that currently only encodings that are a superset of ASCII work, strings of other encodings should be passed in as unicode.

object_hook, if specified, will be called with the result of every JSON object decoded and its return value will be used in place of the given dict. This can be used to provide custom deserializations (e.g. to support JSON-RPC class hinting).

object_pairs_hook, if specified will be called with the result of every JSON object decoded with an ordered list of pairs. The return value of object_pairs_hook will be used instead of the dict. This feature can be used to implement custom decoders that rely on the order that the key and value pairs are decoded (for example, collections.OrderedDict() will remember the order of insertion). If object_hook is also defined, the object_pairs_hook takes priority.

在 2.7 版更改: Added support for object_pairs_hook.

parse_float, if specified, will be called with the string of every JSON float to be decoded. By default, this is equivalent to float(num_str). This can be used to use another datatype or parser for JSON floats (e.g. decimal.Decimal).

parse_int, if specified, will be called with the string of every JSON int to be decoded. By default, this is equivalent to int(num_str). This can be used to use another datatype or parser for JSON integers (e.g. float).

parse_constant, if specified, will be called with one of the following strings: '-Infinity', 'Infinity', 'NaN'. This can be used to raise an exception if invalid JSON numbers are encountered.

If strict is false (True is the default), then control characters will be allowed inside strings. Control characters in this context are those with character codes in the 0–31 range, including '\t' (tab), '\n', '\r' and '\0'.

If the data being deserialized is not a valid JSON document, a ValueError will be raised.

decode(s)

Return the Python representation of s (a str or unicode instance containing a JSON document).

raw_decode(s)

Decode a JSON document from s (a str or unicode beginning with a JSON document) and return a 2-tuple of the Python representation and the index in s where the document ended.

This can be used to decode a JSON document from a string that may have extraneous data at the end.

class json.JSONEncoder([skipkeys[, ensure_ascii[, check_circular[, allow_nan[, sort_keys[, indent[, separators[, encoding[, default]]]]]]]]])

Extensible JSON encoder for Python data structures.

Supports the following objects and types by default:

Python JSON
dict object
list, tuple array
str, unicode string
int, long, float number
True true
False false
None null

To extend this to recognize other objects, subclass and implement a default() method with another method that returns a serializable object for o if possible, otherwise it should call the superclass implementation (to raise TypeError).

If skipkeys is false (the default), then it is a TypeError to attempt encoding of keys that are not str, int, long, float or None. If skipkeys is true, such items are simply skipped.

If ensure_ascii is true (the default), all non-ASCII characters in the output are escaped with \uXXXX sequences, and the results are str instances consisting of ASCII characters only. If ensure_ascii is false, a result may be a unicode instance. This usually happens if the input contains unicode strings or the encoding parameter is used.

If check_circular is true (the default), then lists, dicts, and custom encoded objects will be checked for circular references during encoding to prevent an infinite recursion (which would cause an OverflowError). Otherwise, no such check takes place.

If allow_nan is true (the default), then NaN, Infinity, and -Infinity will be encoded as such. This behavior is not JSON specification compliant, but is consistent with most JavaScript based encoders and decoders. Otherwise, it will be a ValueError to encode such floats.

If sort_keys is true (default: False), then the output of dictionaries will be sorted by key; this is useful for regression tests to ensure that JSON serializations can be compared on a day-to-day basis.

If indent is a non-negative integer (it is None by default), then JSON array elements and object members will be pretty-printed with that indent level. An indent level of 0 will only insert newlines. None is the most compact representation.

注解

Since the default item separator is ', ', the output might include trailing whitespace when indent is specified. You can use separators=(',', ': ') to avoid this.

If specified, separators should be an (item_separator, key_separator) tuple. By default, (', ', ': ') are used. To get the most compact JSON representation, you should specify (',', ':') to eliminate whitespace.

default 被指定时,其应该是一个函数,每当某个对象无法被序列化时它会被调用。它应该返回该对象的一个可以被 JSON 编码的版本或者引发一个 TypeError。如果没有被指定,则会直接引发 TypeError

If encoding is not None, then all input strings will be transformed into unicode using that encoding prior to JSON-encoding. The default is UTF-8.

default(o)

Implement this method in a subclass such that it returns a serializable object for o, or calls the base implementation (to raise a TypeError).

For example, to support arbitrary iterators, you could implement default like this:

def default(self, o):
   try:
       iterable = iter(o)
   except TypeError:
       pass
   else:
       return list(iterable)
   # Let the base class default method raise the TypeError
   return JSONEncoder.default(self, o)
encode(o)

Return a JSON string representation of a Python data structure, o. For example:

>>> JSONEncoder().encode({"foo": ["bar", "baz"]})
'{"foo": ["bar", "baz"]}'
iterencode(o)

Encode the given object, o, and yield each string representation as available. For example:

for chunk in JSONEncoder().iterencode(bigobject):
    mysocket.write(chunk)

18.2.3. Standard Compliance and Interoperability

The JSON format is specified by RFC 7159 and by ECMA-404. This section details this module’s level of compliance with the RFC. For simplicity, JSONEncoder and JSONDecoder subclasses, and parameters other than those explicitly mentioned, are not considered.

This module does not comply with the RFC in a strict fashion, implementing some extensions that are valid JavaScript but not valid JSON. In particular:

  • Infinite and NaN number values are accepted and output;
  • Repeated names within an object are accepted, and only the value of the last name-value pair is used.

Since the RFC permits RFC-compliant parsers to accept input texts that are not RFC-compliant, this module’s deserializer is technically RFC-compliant under default settings.

18.2.3.1. Character Encodings

The RFC requires that JSON be represented using either UTF-8, UTF-16, or UTF-32, with UTF-8 being the recommended default for maximum interoperability. Accordingly, this module uses UTF-8 as the default for its encoding parameter.

This module’s deserializer only directly works with ASCII-compatible encodings; UTF-16, UTF-32, and other ASCII-incompatible encodings require the use of workarounds described in the documentation for the deserializer’s encoding parameter.

As permitted, though not required, by the RFC, this module’s serializer sets ensure_ascii=True by default, thus escaping the output so that the resulting strings only contain ASCII characters.

The RFC prohibits adding a byte order mark (BOM) to the start of a JSON text, and this module’s serializer does not add a BOM to its output. The RFC permits, but does not require, JSON deserializers to ignore an initial BOM in their input. This module’s deserializer raises a ValueError when an initial BOM is present.

The RFC does not explicitly forbid JSON strings which contain byte sequences that don’t correspond to valid Unicode characters (e.g. unpaired UTF-16 surrogates), but it does note that they may cause interoperability problems. By default, this module accepts and outputs (when present in the original str) code points for such sequences.

18.2.3.2. Infinite and NaN Number Values

The RFC does not permit the representation of infinite or NaN number values. Despite that, by default, this module accepts and outputs Infinity, -Infinity, and NaN as if they were valid JSON number literal values:

>>> # Neither of these calls raises an exception, but the results are not valid JSON
>>> json.dumps(float('-inf'))
'-Infinity'
>>> json.dumps(float('nan'))
'NaN'
>>> # Same when deserializing
>>> json.loads('-Infinity')
-inf
>>> json.loads('NaN')
nan

In the serializer, the allow_nan parameter can be used to alter this behavior. In the deserializer, the parse_constant parameter can be used to alter this behavior.

18.2.3.3. Repeated Names Within an Object

The RFC specifies that the names within a JSON object should be unique, but does not mandate how repeated names in JSON objects should be handled. By default, this module does not raise an exception; instead, it ignores all but the last name-value pair for a given name:

>>> weird_json = '{"x": 1, "x": 2, "x": 3}'
>>> json.loads(weird_json)
{u'x': 3}

The object_pairs_hook parameter can be used to alter this behavior.

18.2.3.4. Top-level Non-Object, Non-Array Values

The old version of JSON specified by the obsolete RFC 4627 required that the top-level value of a JSON text must be either a JSON object or array (Python dict or list), and could not be a JSON null, boolean, number, or string value. RFC 7159 removed that restriction, and this module does not and has never implemented that restriction in either its serializer or its deserializer.

Regardless, for maximum interoperability, you may wish to voluntarily adhere to the restriction yourself.

18.2.3.5. Implementation Limitations

Some JSON deserializer implementations may set limits on:

  • the size of accepted JSON texts
  • the maximum level of nesting of JSON objects and arrays
  • the range and precision of JSON numbers
  • the content and maximum length of JSON strings

This module does not impose any such limits beyond those of the relevant Python datatypes themselves or the Python interpreter itself.

When serializing to JSON, beware any such limitations in applications that may consume your JSON. In particular, it is common for JSON numbers to be deserialized into IEEE 754 double precision numbers and thus subject to that representation’s range and precision limitations. This is especially relevant when serializing Python int values of extremely large magnitude, or when serializing instances of “exotic” numerical types such as decimal.Decimal.

脚注

[1]As noted in the errata for RFC 7159, JSON permits literal U+2028 (LINE SEPARATOR) and U+2029 (PARAGRAPH SEPARATOR) characters in strings, whereas JavaScript (as of ECMAScript Edition 5.1) does not.