json
— JSON encoder and decoder¶
Source code:Lib/json/__init__.py
JSON (JavaScript Object Notation), specified byRFC 7159 (which obsoletesRFC 4627) and byECMA-404,is a lightweight data interchange format inspired byJavaScript object literal syntax(although it is not a strict subset of JavaScript[1] ).
Note
The term “object” in the context of JSON processing in Python can beambiguous. All values in Python are objects. In JSON, an object refers toany data wrapped in curly braces, similar to a Python dictionary.
Warning
Be cautious when parsing JSON data from untrusted sources. A maliciousJSON string may cause the decoder to consume considerable CPU and memoryresources. Limiting the size of data to be parsed is recommended.
This module exposes an API familiar to users of the standard librarymarshal
andpickle
modules.
Encoding basic Python object hierarchies:
>>>importjson>>>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('\u1234'))"\u1234">>>print(json.dumps('\\'))"\\">>>print(json.dumps({"c":0,"b":0,"a":0},sort_keys=True)){"a": 0, "b": 0, "c": 0}>>>fromioimportStringIO>>>io=StringIO()>>>json.dump(['streaming API'],io)>>>io.getvalue()'["streaming API"]'
Compact encoding:
>>>importjson>>>json.dumps([1,2,3,{'4':5,'6':7}],separators=(',',':'))'[1,2,3,{"4":5,"6":7}]'
Pretty printing:
>>>importjson>>>print(json.dumps({'6':7,'4':5},sort_keys=True,indent=4)){ "4": 5, "6": 7}
Customizing JSON object encoding:
>>>importjson>>>defcustom_json(obj):...ifisinstance(obj,complex):...return{'__complex__':True,'real':obj.real,'imag':obj.imag}...raiseTypeError(f'Cannot serialize object of{type(obj)}')...>>>json.dumps(1+2j,default=custom_json)'{"__complex__": true, "real": 1.0, "imag": 2.0}'
Decoding JSON:
>>>importjson>>>json.loads('["foo", {"bar":["baz", null, 1.0, 2]}]')['foo', {'bar': ['baz', None, 1.0, 2]}]>>>json.loads('"\\"foo\\bar"')'"foo\x08ar'>>>fromioimportStringIO>>>io=StringIO('["streaming API"]')>>>json.load(io)['streaming API']
Customizing JSON object decoding:
>>>importjson>>>defas_complex(dct):...if'__complex__'indct:...returncomplex(dct['real'],dct['imag'])...returndct...>>>json.loads('{"__complex__": true, "real": 1, "imag": 2}',...object_hook=as_complex)(1+2j)>>>importdecimal>>>json.loads('1.1',parse_float=decimal.Decimal)Decimal('1.1')
ExtendingJSONEncoder
:
>>>importjson>>>classComplexEncoder(json.JSONEncoder):...defdefault(self,obj):...ifisinstance(obj,complex):...return[obj.real,obj.imag]...# Let the base class default method raise the TypeError...returnsuper().default(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', ']']
Usingjson.tool
from the shell to validate and pretty-print:
$echo'{"json":"obj"}'|python-mjson.tool{ "json": "obj"}$echo'{1.2:3.4}'|python-mjson.toolExpecting property name enclosed in double quotes: line 1 column 2 (char 1)
SeeCommand Line Interface for detailed documentation.
Note
JSON is a subset ofYAML 1.2. The JSON produced bythis module’s default settings (in particular, the defaultseparatorsvalue) is also a subset of YAML 1.0 and 1.1. This module can thus also beused as a YAML serializer.
Note
This module’s encoders and decoders preserve input and output order bydefault. Order is only lost if the underlying containers are unordered.
Basic Usage¶
- json.dump(obj,fp,*,skipkeys=False,ensure_ascii=True,check_circular=True,allow_nan=True,cls=None,indent=None,separators=None,default=None,sort_keys=False,**kw)¶
Serializeobj as a JSON formatted stream tofp (a
.write()
-supportingfile-like object) using thisPython-to-JSON conversion table.Note
Unlike
pickle
andmarshal
, JSON is not a framed protocol,so trying to serialize multiple objects with repeated calls todump()
using the samefp will result in an invalid JSON file.- Parameters:
obj (object) – The Python object to be serialized.
fp (file-like object) – The file-like objectobj will be serialized to.The
json
module always producesstr
objects,notbytes
objects,thereforefp.write()
must supportstr
input.skipkeys (bool) – If
True
, keys that are not of a basic type(str
,int
,float
,bool
,None
)will be skipped instead of raising aTypeError
.DefaultFalse
.ensure_ascii (bool) – If
True
(the default), the output is guaranteed tohave all incoming non-ASCII characters escaped.IfFalse
, these characters will be outputted as-is.check_circular (bool) – If
False
, the circular reference check for container types is skippedand a circular reference will result in aRecursionError
(or worse).DefaultTrue
.allow_nan (bool) – If
False
, serialization of out-of-rangefloat
values(nan
,inf
,-inf
) will result in aValueError
,in strict compliance with the JSON specification.IfTrue
(the default), their JavaScript equivalents(NaN
,Infinity
,-Infinity
) are used.cls (a
JSONEncoder
subclass) – If set, a custom JSON encoder with thedefault()
method overridden,for serializing into custom datatypes.IfNone
(the default),JSONEncoder
is used.indent (int |str |None) – If a positive integer or string, JSON array elements andobject members will be pretty-printed with that indent level.A positive integer indents that many spaces per level;a string (such as
"\t"
) is used to indent each level.If zero, negative, or""
(the empty string),only newlines are inserted.IfNone
(the default), the most compact representation is used.separators (tuple |None) – A two-tuple:
(item_separator,key_separator)
.IfNone
(the default),separators defaults to(',',':')
ifindent isNone
,and(',',':')
otherwise.For the most compact JSON,specify(',',':')
to eliminate whitespace.default (callable | None) – A function that is called for objects that can’t otherwise be serialized.It should return a JSON encodable version of the objector raise a
TypeError
.IfNone
(the default),TypeError
is raised.sort_keys (bool) – If
True
, dictionaries will be outputted sorted by key.DefaultFalse
.
Changed in version 3.2:Allow strings forindent in addition to integers.
Changed in version 3.4:Use
(',',':')
as default ifindent is notNone
.Changed in version 3.6:All optional parameters are nowkeyword-only.
- json.dumps(obj,*,skipkeys=False,ensure_ascii=True,check_circular=True,allow_nan=True,cls=None,indent=None,separators=None,default=None,sort_keys=False,**kw)¶
Serializeobj to a JSON formatted
str
using thisconversiontable. The arguments have the same meaning as indump()
.Note
Keys in key/value pairs of JSON are always of the type
str
. Whena dictionary is converted into JSON, all the keys of the dictionary arecoerced to strings. As a result of this, if a dictionary is convertedinto JSON and then back into a dictionary, the dictionary may not equalthe original one. That is,loads(dumps(x))!=x
if x has non-stringkeys.
- json.load(fp,*,cls=None,object_hook=None,parse_float=None,parse_int=None,parse_constant=None,object_pairs_hook=None,**kw)¶
Deserializefp to a Python objectusing theJSON-to-Python conversion table.
- Parameters:
fp (file-like object) – A
.read()
-supportingtext file orbinary filecontaining the JSON document to be deserialized.cls (a
JSONDecoder
subclass) – If set, a custom JSON decoder.Additional keyword arguments toload()
will be passed to the constructor ofcls.IfNone
(the default),JSONDecoder
is used.object_hook (callable | None) – If set, a function that is called with the result ofany JSON object literal decoded (a
dict
).The return value of this function will be usedinstead of thedict
.This feature can be used to implement custom decoders,for exampleJSON-RPC class hinting.DefaultNone
.object_pairs_hook (callable | None) – If set, a function that is called with the result ofany JSON object literal decoded with an ordered list of pairs.The return value of this function will be usedinstead of the
dict
.This feature can be used to implement custom decoders.Ifobject_hook is also set,object_pairs_hook takes priority.DefaultNone
.parse_float (callable | None) – If set, a function that is called withthe string of every JSON float to be decoded.If
None
(the default), it is equivalent tofloat(num_str)
.This can be used to parse JSON floats into custom datatypes,for exampledecimal.Decimal
.parse_int (callable | None) – If set, a function that is called withthe string of every JSON int to be decoded.If
None
(the default), it is equivalent toint(num_str)
.This can be used to parse JSON integers into custom datatypes,for examplefloat
.parse_constant (callable | None) – If set, a function that is called with one of the following strings:
'-Infinity'
,'Infinity'
, or'NaN'
.This can be used to raise an exceptionif invalid JSON numbers are encountered.DefaultNone
.
- Raises:
JSONDecodeError – When the data being deserialized is not a valid JSON document.
UnicodeDecodeError – When the data being deserialized does not containUTF-8, UTF-16 or UTF-32 encoded data.
Changed in version 3.1:
Added the optionalobject_pairs_hook parameter.
parse_constant doesn’t get called on ‘null’, ‘true’, ‘false’ anymore.
Changed in version 3.6:
All optional parameters are nowkeyword-only.
fp can now be abinary file.The input encoding should be UTF-8, UTF-16 or UTF-32.
Changed in version 3.11:The defaultparse_int of
int()
now limits the maximum length ofthe integer string via the interpreter’sinteger stringconversion length limitation to help avoid denialof service attacks.
- json.loads(s,*,cls=None,object_hook=None,parse_float=None,parse_int=None,parse_constant=None,object_pairs_hook=None,**kw)¶
Identical to
load()
, but instead of a file-like object,deserializes (astr
,bytes
orbytearray
instance containing a JSON document) to a Python object using thisconversion table.Changed in version 3.6:s can now be of type
bytes
orbytearray
. Theinput encoding should be UTF-8, UTF-16 or UTF-32.Changed in version 3.9:The keyword argumentencoding has been removed.
Encoders and Decoders¶
- classjson.JSONDecoder(*,object_hook=None,parse_float=None,parse_int=None,parse_constant=None,strict=True,object_pairs_hook=None)¶
Simple JSON decoder.
Performs the following translations in decoding by default:
JSON
Python
object
dict
array
list
string
str
number (int)
int
number (real)
float
true
True
false
False
null
None
It also understands
NaN
,Infinity
, and-Infinity
as theircorrespondingfloat
values, which is outside the JSON spec.object_hook is an optional function that will be called with the result ofevery JSON object decoded and its return value will be used in place of thegiven
dict
. This can be used to provide custom deserializations(e.g. to supportJSON-RPC class hinting).object_pairs_hook is an optional function that will be called with theresult of every JSON object decoded with an ordered list of pairs. Thereturn value ofobject_pairs_hook will be used instead of the
dict
. This feature can be used to implement custom decoders. Ifobject_hook is also defined, theobject_pairs_hook takes priority.Changed in version 3.1:Added support forobject_pairs_hook.
parse_float is an optional function that will be called with the string ofevery JSON float to be decoded. By default, this is equivalent to
float(num_str)
. This can be used to use another datatype or parser forJSON floats (e.g.decimal.Decimal
).parse_int is an optional function that will be called with the string ofevery JSON int to be decoded. By default, this is equivalent to
int(num_str)
. This can be used to use another datatype or parser forJSON integers (e.g.float
).parse_constant is an optional function that will be called with one of thefollowing strings:
'-Infinity'
,'Infinity'
,'NaN'
. This can beused to raise an exception if invalid JSON numbers are encountered.Ifstrict is false (
True
is the default), then control characterswill be allowed inside strings. Control characters in this context arethose 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
JSONDecodeError
will be raised.Changed in version 3.6:All parameters are nowkeyword-only.
- decode(s)¶
Return the Python representation ofs (a
str
instancecontaining a JSON document).JSONDecodeError
will be raised if the given JSON document is notvalid.
- classjson.JSONEncoder(*,skipkeys=False,ensure_ascii=True,check_circular=True,allow_nan=True,sort_keys=False,indent=None,separators=None,default=None)¶
Extensible JSON encoder for Python data structures.
Supports the following objects and types by default:
Python
JSON
dict
object
list, tuple
array
str
string
int, float, int- & float-derived Enums
number
True
true
False
false
None
null
Changed in version 3.4:Added support for int- and float-derived Enum classes.
To extend this to recognize other objects, subclass and implement a
default()
method with another method that returns a serializable objectforo
if possible, otherwise it should call the superclass implementation(to raiseTypeError
).Ifskipkeys is false (the default), a
TypeError
will be raised whentrying to encode keys that are notstr
,int
,float
,bool
orNone
. Ifskipkeys is true, such items are simply skipped.Ifensure_ascii is true (the default), the output is guaranteed tohave all incoming non-ASCII characters escaped. Ifensure_ascii isfalse, these characters will be output as-is.
Ifcheck_circular is true (the default), then lists, dicts, and customencoded objects will be checked for circular references during encoding toprevent an infinite recursion (which would cause a
RecursionError
).Otherwise, no such check takes place.Ifallow_nan is true (the default), then
NaN
,Infinity
, and-Infinity
will be encoded as such. This behavior is not JSONspecification compliant, but is consistent with most JavaScript basedencoders and decoders. Otherwise, it will be aValueError
to encodesuch floats.Ifsort_keys is true (default:
False
), then the output of dictionarieswill be sorted by key; this is useful for regression tests to ensure thatJSON serializations can be compared on a day-to-day basis.Ifindent is a non-negative integer or string, then JSON array elements andobject members will be pretty-printed with that indent level. An indent levelof 0, negative, or
""
will only insert newlines.None
(the default)selects the most compact representation. Using a positive integer indentindents that many spaces per level. Ifindent is a string (such as"\t"
),that string is used to indent each level.Changed in version 3.2:Allow strings forindent in addition to integers.
If specified,separators should be an
(item_separator,key_separator)
tuple. The default is(',',':')
ifindent isNone
and(',',':')
otherwise. To get the most compact JSON representation,you should specify(',',':')
to eliminate whitespace.Changed in version 3.4:Use
(',',':')
as default ifindent is notNone
.If specified,default should be a function that gets called for objects thatcan’t otherwise be serialized. It should return a JSON encodable version ofthe object or raise a
TypeError
. If not specified,TypeError
is raised.Changed in version 3.6:All parameters are nowkeyword-only.
- default(o)¶
Implement this method in a subclass such that it returns a serializableobject foro, or calls the base implementation (to raise a
TypeError
).For example, to support arbitrary iterators, you could implement
default()
like this:defdefault(self,o):try:iterable=iter(o)exceptTypeError:passelse:returnlist(iterable)# Let the base class default method raise the TypeErrorreturnsuper().default(o)
- encode(o)¶
Return a JSON string representation of a Python data structure,o. Forexample:
>>>json.JSONEncoder().encode({"foo":["bar","baz"]})'{"foo": ["bar", "baz"]}'
- iterencode(o)¶
Encode the given object,o, and yield each string representation asavailable. For example:
forchunkinjson.JSONEncoder().iterencode(bigobject):mysocket.write(chunk)
Exceptions¶
- exceptionjson.JSONDecodeError(msg,doc,pos)¶
Subclass of
ValueError
with the following additional attributes:- msg¶
The unformatted error message.
- doc¶
The JSON document being parsed.
- pos¶
The start index ofdoc where parsing failed.
- lineno¶
The line corresponding topos.
- colno¶
The column corresponding topos.
Added in version 3.5.
Standard Compliance and Interoperability¶
The JSON format is specified byRFC 7159 and byECMA-404.This section details this module’s level of compliance with the RFC.For simplicity,JSONEncoder
andJSONDecoder
subclasses, andparameters other than those explicitly mentioned, are not considered.
This module does not comply with the RFC in a strict fashion, implementing someextensions 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 lastname-value pair is used.
Since the RFC permits RFC-compliant parsers to accept input texts that are notRFC-compliant, this module’s deserializer is technically RFC-compliant underdefault settings.
Character Encodings¶
The RFC requires that JSON be represented using either UTF-8, UTF-16, orUTF-32, with UTF-8 being the recommended default for maximum interoperability.
As permitted, though not required, by the RFC, this module’s serializer setsensure_ascii=True by default, thus escaping the output so that the resultingstrings only contain ASCII characters.
Other than theensure_ascii parameter, this module is defined strictly interms of conversion between Python objects andUnicodestrings
, and thus does not otherwise directly addressthe issue of character encodings.
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 initialBOM in their input. This module’s deserializer raises aValueError
when an initial BOM is present.
The RFC does not explicitly forbid JSON strings which contain byte sequencesthat don’t correspond to valid Unicode characters (e.g. unpaired UTF-16surrogates), but it does note that they may cause interoperability problems.By default, this module accepts and outputs (when present in the originalstr
) code points for such sequences.
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 outputsInfinity
,-Infinity
, andNaN
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, theallow_nan parameter can be used to alter thisbehavior. In the deserializer, theparse_constant parameter can be used toalter this behavior.
Repeated Names Within an Object¶
The RFC specifies that the names within a JSON object should be unique, butdoes not mandate how repeated names in JSON objects should be handled. Bydefault, this module does not raise an exception; instead, it ignores all butthe last name-value pair for a given name:
>>>weird_json='{"x": 1, "x": 2, "x": 3}'>>>json.loads(weird_json){'x': 3}
Theobject_pairs_hook parameter can be used to alter this behavior.
Top-level Non-Object, Non-Array Values¶
The old version of JSON specified by the obsoleteRFC 4627 required thatthe top-level value of a JSON text must be either a JSON object or array(Pythondict
orlist
), and could not be a JSON null,boolean, number, or string value.RFC 7159 removed that restriction, andthis module does not and has never implemented that restriction in either itsserializer or its deserializer.
Regardless, for maximum interoperability, you may wish to voluntarily adhereto the restriction yourself.
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 relevantPython datatypes themselves or the Python interpreter itself.
When serializing to JSON, beware any such limitations in applications that mayconsume your JSON. In particular, it is common for JSON numbers to bedeserialized into IEEE 754 double precision numbers and thus subject to thatrepresentation’s range and precision limitations. This is especially relevantwhen serializing Pythonint
values of extremely large magnitude, orwhen serializing instances of “exotic” numerical types such asdecimal.Decimal
.
Command Line Interface¶
Source code:Lib/json/tool.py
Thejson.tool
module provides a simple command line interface to validateand pretty-print JSON objects.
If the optionalinfile
andoutfile
arguments are notspecified,sys.stdin
andsys.stdout
will be used respectively:
$echo'{"json": "obj"}'|python-mjson.tool{ "json": "obj"}$echo'{1.2:3.4}'|python-mjson.toolExpecting property name enclosed in double quotes: line 1 column 2 (char 1)
Changed in version 3.5:The output is now in the same order as the input. Use the--sort-keys
option to sort the output of dictionariesalphabetically by key.
Command line options¶
- infile¶
The JSON file to be validated or pretty-printed:
$python-mjson.toolmp_films.json[ { "title": "And Now for Something Completely Different", "year": 1971 }, { "title": "Monty Python and the Holy Grail", "year": 1975 }]
Ifinfile is not specified, read from
sys.stdin
.
- outfile¶
Write the output of theinfile to the givenoutfile. Otherwise, write itto
sys.stdout
.
- --sort-keys¶
Sort the output of dictionaries alphabetically by key.
Added in version 3.5.
- --no-ensure-ascii¶
Disable escaping of non-ascii characters, see
json.dumps()
for more information.Added in version 3.9.
- --json-lines¶
Parse every input line as separate JSON object.
Added in version 3.8.
- --indent,--tab,--no-indent,--compact¶
Mutually exclusive options for whitespace control.
Added in version 3.9.
- -h,--help¶
Show the help message.
Footnotes
[1]As noted inthe errata for RFC 7159,JSON permits literal U+2028 (LINE SEPARATOR) andU+2029 (PARAGRAPH SEPARATOR) characters in strings, whereas JavaScript(as of ECMAScript Edition 5.1) does not.