Python 实用宝典

Question 1

I’m used to bringing data in and out of Python using CSV files, but there are obvious challenges to this. Are there simple ways to store a dictionary (or sets of dictionaries) in a JSON or pickle file?

For example:

data = {}
data ['key1'] = "keyinfo"
data ['key2'] = "keyinfo2"

I would like to know both how to save this, and then how to load it back in.

Question 2

Pickle save:

try:
    import cPickle as pickle
except ImportError:  # Python 3.x
    import pickle

with open('data.p', 'wb') as fp:
    pickle.dump(data, fp, protocol=pickle.HIGHEST_PROTOCOL)

See the pickle module documentation for additional information regarding the protocol argument.

Pickle load:

with open('data.p', 'rb') as fp:
    data = pickle.load(fp)

JSON save:

import json

with open('data.json', 'w') as fp:
    json.dump(data, fp)

Supply extra arguments, like sort_keys or indent, to get a pretty result. The argument sort_keys will sort the keys alphabetically and indent will indent your data structure with indent=N spaces.

json.dump(data, fp, sort_keys=True, indent=4)

JSON load:

with open('data.json', 'r') as fp:
    data = json.load(fp)

Question 3

Minimal example, writing directly to a file:

import json
json.dump(data, open(filename, 'wb'))
data = json.load(open(filename))

or safely opening / closing:

import json
with open(filename, 'wb') as outfile:
    json.dump(data, outfile)
with open(filename) as infile:
    data = json.load(infile)

If you want to save it in a string instead of a file:

import json
json_str = json.dumps(data)
data = json.loads(json_str)

Question 4

Also see the speeded-up package ujson:

import ujson

with open('data.json', 'wb') as fp:
    ujson.dump(data, fp)

Question 5

To write to a file:

import json
myfile.write(json.dumps(mydict))

To read from a file:

import json
mydict = json.loads(myfile.read())

myfile is the file object for the file that you stored the dict in.

Question 6

If you’re after serialization, but won’t need the data in other programs, I strongly recommend the shelve module. Think of it as a persistent dictionary.

myData = shelve.open('/path/to/file')

# Check for values.
keyVar in myData

# Set values
myData[anotherKey] = someValue

# Save the data for future use.
myData.close()

Question 7

If you want an alternative to pickle or json, you can use klepto.

>>> init = {'y': 2, 'x': 1, 'z': 3}
>>> import klepto
>>> cache = klepto.archives.file_archive('memo', init, serialized=False)
>>> cache        
{'y': 2, 'x': 1, 'z': 3}
>>>
>>> # dump dictionary to the file 'memo.py'
>>> cache.dump() 
>>> 
>>> # import from 'memo.py'
>>> from memo import memo
>>> print memo
{'y': 2, 'x': 1, 'z': 3}

With klepto, if you had used serialized=True, the dictionary would have been written to memo.pkl as a pickled dictionary instead of with clear text.

You can get klepto here: https://github.com/uqfoundation/klepto

dill is probably a better choice for pickling then pickle itself, as dill can serialize almost anything in python. klepto also can use dill.

You can get dill here: https://github.com/uqfoundation/dill

The additional mumbo-jumbo on the first few lines are because klepto can be configured to store dictionaries to a file, to a directory context, or to a SQL database. The API is the same for whatever you choose as the backend archive. It gives you an “archivable” dictionary with which you can use load and dump to interact with the archive.

Question 8

For completeness, we should include ConfigParser and configparser which are part of the standard library in Python 2 and 3, respectively. This module reads and writes to a config/ini file and (at least in Python 3) behaves in a lot of ways like a dictionary. It has the added benefit that you can store multiple dictionaries into separate sections of your config/ini file and recall them. Sweet!

Python 2.7.x example.

import ConfigParser

config = ConfigParser.ConfigParser()

dict1 = {'key1':'keyinfo', 'key2':'keyinfo2'}
dict2 = {'k1':'hot', 'k2':'cross', 'k3':'buns'}
dict3 = {'x':1, 'y':2, 'z':3}

# Make each dictionary a separate section in the configuration
config.add_section('dict1')
for key in dict1.keys():
    config.set('dict1', key, dict1[key])
   
config.add_section('dict2')
for key in dict2.keys():
    config.set('dict2', key, dict2[key])

config.add_section('dict3')
for key in dict3.keys():
    config.set('dict3', key, dict3[key])

# Save the configuration to a file
f = open('config.ini', 'w')
config.write(f)
f.close()

# Read the configuration from a file
config2 = ConfigParser.ConfigParser()
config2.read('config.ini')

dictA = {}
for item in config2.items('dict1'):
    dictA[item[0]] = item[1]

dictB = {}
for item in config2.items('dict2'):
    dictB[item[0]] = item[1]

dictC = {}
for item in config2.items('dict3'):
    dictC[item[0]] = item[1]

print(dictA)
print(dictB)
print(dictC)

Python 3.X example.

import configparser

config = configparser.ConfigParser()

dict1 = {'key1':'keyinfo', 'key2':'keyinfo2'}
dict2 = {'k1':'hot', 'k2':'cross', 'k3':'buns'}
dict3 = {'x':1, 'y':2, 'z':3}

# Make each dictionary a separate section in the configuration
config['dict1'] = dict1
config['dict2'] = dict2
config['dict3'] = dict3

# Save the configuration to a file
f = open('config.ini', 'w')
config.write(f)
f.close()

# Read the configuration from a file
config2 = configparser.ConfigParser()
config2.read('config.ini')

# ConfigParser objects are a lot like dictionaries, but if you really
# want a dictionary you can ask it to convert a section to a dictionary
dictA = dict(config2['dict1'] )
dictB = dict(config2['dict2'] )
dictC = dict(config2['dict3'])

print(dictA)
print(dictB)
print(dictC)

Console output

{'key2': 'keyinfo2', 'key1': 'keyinfo'}
{'k1': 'hot', 'k2': 'cross', 'k3': 'buns'}
{'z': '3', 'y': '2', 'x': '1'}

Contents of config.ini

[dict1]
key2 = keyinfo2
key1 = keyinfo

[dict2]
k1 = hot
k2 = cross
k3 = buns

[dict3]
z = 3
y = 2
x = 1

Question 9

If save to a JSON file, the best and easiest way of doing this is:

import json
with open("file.json", "wb") as f:
    f.write(json.dumps(dict).encode("utf-8"))

Question 10

My use case was to save multiple JSON objects to a file and marty’s answer helped me somewhat. But to serve my use case, the answer was not complete as it would overwrite the old data every time a new entry was saved.

To save multiple entries in a file, one must check for the old content (i.e., read before write). A typical file holding JSON data will either have a list or an object as root. So I considered that my JSON file always has a list of objects and every time I add data to it, I simply load the list first, append my new data in it, and dump it back to a writable-only instance of file (w):

def saveJson(url,sc): # This function writes the two values to the file
    newdata = {'url':url,'sc':sc}
    json_path = "db/file.json"

    old_list= []
    with open(json_path) as myfile:  # Read the contents first
        old_list = json.load(myfile)
    old_list.append(newdata)

    with open(json_path,"w") as myfile:  # Overwrite the whole content
        json.dump(old_list, myfile, sort_keys=True, indent=4)

    return "success"

The new JSON file will look something like this:

[
    {
        "sc": "a11",
        "url": "www.google.com"
    },
    {
        "sc": "a12",
        "url": "www.google.com"
    },
    {
        "sc": "a13",
        "url": "www.google.com"
    }
]

NOTE: It is essential to have a file named file.json with [] as initial data for this approach to work

PS: not related to original question, but this approach could also be further improved by first checking if our entry already exists (based on one or multiple keys) and only then append and save the data.

Question 11

I am trying to create a JSON string representation of a class instance and having difficulty. Let’s say the class is built like this:

class testclass:
    value1 = "a"
    value2 = "b"

A call to the json.dumps is made like this:

t = testclass()
json.dumps(t)

It is failing and telling me that the testclass is not JSON serializable.

TypeError: <__main__.testclass object at 0x000000000227A400> is not JSON serializable

I have also tried using the pickle module :

t = testclass()
print(pickle.dumps(t, pickle.HIGHEST_PROTOCOL))

And it gives class instance information but not a serialized content of the class instance.

b'\x80\x03c__main__\ntestclass\nq\x00)\x81q\x01}q\x02b.'

What am I doing wrong?

Question 12

The basic problem is that the JSON encoder json.dumps() only knows how to serialize a limited set of object types by default, all built-in types. List here: https://docs.python.org/3.3/library/json.html#encoders-and-decoders

One good solution would be to make your class inherit from JSONEncoder and then implement the JSONEncoder.default() function, and make that function emit the correct JSON for your class.

A simple solution would be to call json.dumps() on the .__dict__ member of that instance. That is a standard Python dict and if your class is simple it will be JSON serializable.

class Foo(object):
    def __init__(self):
        self.x = 1
        self.y = 2

foo = Foo()
s = json.dumps(foo) # raises TypeError with "is not JSON serializable"

s = json.dumps(foo.__dict__) # s set to: {"x":1, "y":2}

The above approach is discussed in this blog posting:

Serializing arbitrary Python objects to JSON using __dict__

Question 13

There’s one way that works great for me that you can try out:

json.dumps() can take an optional parameter default where you can specify a custom serializer function for unknown types, which in my case looks like

def serialize(obj):
    """JSON serializer for objects not serializable by default json code"""

    if isinstance(obj, date):
        serial = obj.isoformat()
        return serial

    if isinstance(obj, time):
        serial = obj.isoformat()
        return serial

    return obj.__dict__

First two ifs are for date and time serialization and then there is a obj.__dict__ returned for any other object.

the final call looks like:

json.dumps(myObj, default=serialize)

It’s especially good when you are serializing a collection and you don’t want to call __dict__ explicitly for every object. Here it’s done for you automatically.

So far worked so good for me, looking forward for your thoughts.

Question 14

You can specify the default named parameter in the json.dumps() function:

json.dumps(obj, default=lambda x: x.__dict__)

Explanation:

Form the docs (2.7, 3.6):

``default(obj)`` is a function that should return a serializable version
of obj or raise TypeError. The default simply raises TypeError.

(Works on Python 2.7 and Python 3.x)

Note: In this case you need instance variables and not class variables, as the example in the question tries to do. (I am assuming the asker meant class instance to be an object of a class)

I learned this first from @phihag’s answer here. Found it to be the simplest and cleanest way to do the job.

Question 15

I just do:

data=json.dumps(myobject.__dict__)

This is not the full answer, and if you have some sort of complicated object class you certainly will not get everything. However I use this for some of my simple objects.

One that it works really well on is the “options” class that you get from the OptionParser module. Here it is along with the JSON request itself.

  def executeJson(self, url, options):
        data=json.dumps(options.__dict__)
        if options.verbose:
            print data
        headers = {'Content-type': 'application/json', 'Accept': 'text/plain'}
        return requests.post(url, data, headers=headers)

Question 16

Using jsonpickle

import jsonpickle

object = YourClass()
json_object = jsonpickle.encode(object)

Question 17

JSON is not really meant for serializing arbitrary Python objects. It’s great for serializing dict objects, but the pickle module is really what you should be using in general. Output from pickle is not really human-readable, but it should unpickle just fine. If you insist on using JSON, you could check out the jsonpickle module, which is an interesting hybrid approach.

https://github.com/jsonpickle/jsonpickle

Question 18

Here are two simple functions for serialization of any non-sophisticated classes, nothing fancy as explained before.

I use this for configuration type stuff because I can add new members to the classes with no code adjustments.

import json

class SimpleClass:
    def __init__(self, a=None, b=None, c=None):
        self.a = a
        self.b = b
        self.c = c

def serialize_json(instance=None, path=None):
    dt = {}
    dt.update(vars(instance))

    with open(path, "w") as file:
        json.dump(dt, file)

def deserialize_json(cls=None, path=None):
    def read_json(_path):
        with open(_path, "r") as file:
            return json.load(file)

    data = read_json(path)

    instance = object.__new__(cls)

    for key, value in data.items():
        setattr(instance, key, value)

    return instance

# Usage: Create class and serialize under Windows file system.
write_settings = SimpleClass(a=1, b=2, c=3)
serialize_json(write_settings, r"c:\temp\test.json")

# Read back and rehydrate.
read_settings = deserialize_json(SimpleClass, r"c:\temp\test.json")

# results are the same.
print(vars(write_settings))
print(vars(read_settings))

# output:
# {'c': 3, 'b': 2, 'a': 1}
# {'c': 3, 'b': 2, 'a': 1}

Question 19

There are some good answers on how to get started on doing this. But there are some things to keep in mind:

What if the instance is nested inside a large data structure?
What if also want the class name?
What if you want to deserialize the instance?
What if you’re using __slots__ instead of __dict__?
What if you just don’t want to do it yourself?

json-tricks is a library (that I made and others contributed to) which has been able to do this for quite a while. For example:

class MyTestCls:
    def __init__(self, **kwargs):
        for k, v in kwargs.items():
            setattr(self, k, v)

cls_instance = MyTestCls(s='ub', dct={'7': 7})

json = dumps(cls_instance, indent=4)
instance = loads(json)

You’ll get your instance back. Here the json looks like this:

{
    "__instance_type__": [
        "json_tricks.test_class",
        "MyTestCls"
    ],
    "attributes": {
        "s": "ub",
        "dct": {
            "7": 7
        }
    }
}

If you like to make your own solution, you might look at the source of json-tricks so as not to forget some special cases (like __slots__).

It also does other types like numpy arrays, datetimes, complex numbers; it also allows for comments.

Question 20

Python3.x

The best aproach I could reach with my knowledge was this.
Note that this code treat set() too.
This approach is generic just needing the extension of class (in the second example).
Note that I’m just doing it to files, but it’s easy to modify the behavior to your taste.

However this is a CoDec.

With a little more work you can construct your class in other ways. I assume a default constructor to instance it, then I update the class dict.

import json
import collections


class JsonClassSerializable(json.JSONEncoder):

    REGISTERED_CLASS = {}

    def register(ctype):
        JsonClassSerializable.REGISTERED_CLASS[ctype.__name__] = ctype

    def default(self, obj):
        if isinstance(obj, collections.Set):
            return dict(_set_object=list(obj))
        if isinstance(obj, JsonClassSerializable):
            jclass = {}
            jclass["name"] = type(obj).__name__
            jclass["dict"] = obj.__dict__
            return dict(_class_object=jclass)
        else:
            return json.JSONEncoder.default(self, obj)

    def json_to_class(self, dct):
        if '_set_object' in dct:
            return set(dct['_set_object'])
        elif '_class_object' in dct:
            cclass = dct['_class_object']
            cclass_name = cclass["name"]
            if cclass_name not in self.REGISTERED_CLASS:
                raise RuntimeError(
                    "Class {} not registered in JSON Parser"
                    .format(cclass["name"])
                )
            instance = self.REGISTERED_CLASS[cclass_name]()
            instance.__dict__ = cclass["dict"]
            return instance
        return dct

    def encode_(self, file):
        with open(file, 'w') as outfile:
            json.dump(
                self.__dict__, outfile,
                cls=JsonClassSerializable,
                indent=4,
                sort_keys=True
            )

    def decode_(self, file):
        try:
            with open(file, 'r') as infile:
                self.__dict__ = json.load(
                    infile,
                    object_hook=self.json_to_class
                )
        except FileNotFoundError:
            print("Persistence load failed "
                  "'{}' do not exists".format(file)
                  )


class C(JsonClassSerializable):

    def __init__(self):
        self.mill = "s"


JsonClassSerializable.register(C)


class B(JsonClassSerializable):

    def __init__(self):
        self.a = 1230
        self.c = C()


JsonClassSerializable.register(B)


class A(JsonClassSerializable):

    def __init__(self):
        self.a = 1
        self.b = {1, 2}
        self.c = B()

JsonClassSerializable.register(A)

A().encode_("test")
b = A()
b.decode_("test")
print(b.a)
print(b.b)
print(b.c.a)

Edit

With some more of research I found a way to generalize without the need of the SUPERCLASS register method call, using a metaclass

import json
import collections

REGISTERED_CLASS = {}

class MetaSerializable(type):

    def __call__(cls, *args, **kwargs):
        if cls.__name__ not in REGISTERED_CLASS:
            REGISTERED_CLASS[cls.__name__] = cls
        return super(MetaSerializable, cls).__call__(*args, **kwargs)


class JsonClassSerializable(json.JSONEncoder, metaclass=MetaSerializable):

    def default(self, obj):
        if isinstance(obj, collections.Set):
            return dict(_set_object=list(obj))
        if isinstance(obj, JsonClassSerializable):
            jclass = {}
            jclass["name"] = type(obj).__name__
            jclass["dict"] = obj.__dict__
            return dict(_class_object=jclass)
        else:
            return json.JSONEncoder.default(self, obj)

    def json_to_class(self, dct):
        if '_set_object' in dct:
            return set(dct['_set_object'])
        elif '_class_object' in dct:
            cclass = dct['_class_object']
            cclass_name = cclass["name"]
            if cclass_name not in REGISTERED_CLASS:
                raise RuntimeError(
                    "Class {} not registered in JSON Parser"
                    .format(cclass["name"])
                )
            instance = REGISTERED_CLASS[cclass_name]()
            instance.__dict__ = cclass["dict"]
            return instance
        return dct

    def encode_(self, file):
        with open(file, 'w') as outfile:
            json.dump(
                self.__dict__, outfile,
                cls=JsonClassSerializable,
                indent=4,
                sort_keys=True
            )

    def decode_(self, file):
        try:
            with open(file, 'r') as infile:
                self.__dict__ = json.load(
                    infile,
                    object_hook=self.json_to_class
                )
        except FileNotFoundError:
            print("Persistence load failed "
                  "'{}' do not exists".format(file)
                  )


class C(JsonClassSerializable):

    def __init__(self):
        self.mill = "s"


class B(JsonClassSerializable):

    def __init__(self):
        self.a = 1230
        self.c = C()


class A(JsonClassSerializable):

    def __init__(self):
        self.a = 1
        self.b = {1, 2}
        self.c = B()


A().encode_("test")
b = A()
b.decode_("test")
print(b.a)
# 1
print(b.b)
# {1, 2}
print(b.c.a)
# 1230
print(b.c.c.mill)
# s

Question 21

I believe instead of inheritance as suggested in accepted answer, it’s better to use polymorphism. Otherwise you have to have a big if else statement to customize encoding of every object. That means create a generic default encoder for JSON as:

def jsonDefEncoder(obj):
   if hasattr(obj, 'jsonEnc'):
      return obj.jsonEnc()
   else: #some default behavior
      return obj.__dict__

and then have a jsonEnc() function in each class you want to serialize. e.g.

class A(object):
   def __init__(self,lengthInFeet):
      self.lengthInFeet=lengthInFeet
   def jsonEnc(self):
      return {'lengthInMeters': lengthInFeet * 0.3 } # each foot is 0.3 meter

Then you call json.dumps(classInstance,default=jsonDefEncoder)

Question 22

When I run something like:

from multiprocessing import Pool

p = Pool(5)
def f(x):
     return x*x

p.map(f, [1,2,3])

it works fine. However, putting this as a function of a class:

class calculate(object):
    def run(self):
        def f(x):
            return x*x

        p = Pool()
        return p.map(f, [1,2,3])

cl = calculate()
print cl.run()

Gives me the following error:

Exception in thread Thread-1:
Traceback (most recent call last):
  File "/sw/lib/python2.6/threading.py", line 532, in __bootstrap_inner
    self.run()
  File "/sw/lib/python2.6/threading.py", line 484, in run
    self.__target(*self.__args, **self.__kwargs)
  File "/sw/lib/python2.6/multiprocessing/pool.py", line 225, in _handle_tasks
    put(task)
PicklingError: Can't pickle <type 'function'>: attribute lookup __builtin__.function failed

I’ve seen a post from Alex Martelli dealing with the same kind of problem, but it wasn’t explicit enough.

Question 23

I also was annoyed by restrictions on what sort of functions pool.map could accept. I wrote the following to circumvent this. It appears to work, even for recursive use of parmap.

from multiprocessing import Process, Pipe
from itertools import izip

def spawn(f):
    def fun(pipe, x):
        pipe.send(f(x))
        pipe.close()
    return fun

def parmap(f, X):
    pipe = [Pipe() for x in X]
    proc = [Process(target=spawn(f), args=(c, x)) for x, (p, c) in izip(X, pipe)]
    [p.start() for p in proc]
    [p.join() for p in proc]
    return [p.recv() for (p, c) in pipe]

if __name__ == '__main__':
    print parmap(lambda x: x**x, range(1, 5))

Question 24

I could not use the codes posted so far because the codes using “multiprocessing.Pool” do not work with lambda expressions and the codes not using “multiprocessing.Pool” spawn as many processes as there are work items.

I adapted the code s.t. it spawns a predefined amount of workers and only iterates through the input list if there exists an idle worker. I also enabled the “daemon” mode for the workers s.t. ctrl-c works as expected.

import multiprocessing


def fun(f, q_in, q_out):
    while True:
        i, x = q_in.get()
        if i is None:
            break
        q_out.put((i, f(x)))


def parmap(f, X, nprocs=multiprocessing.cpu_count()):
    q_in = multiprocessing.Queue(1)
    q_out = multiprocessing.Queue()

    proc = [multiprocessing.Process(target=fun, args=(f, q_in, q_out))
            for _ in range(nprocs)]
    for p in proc:
        p.daemon = True
        p.start()

    sent = [q_in.put((i, x)) for i, x in enumerate(X)]
    [q_in.put((None, None)) for _ in range(nprocs)]
    res = [q_out.get() for _ in range(len(sent))]

    [p.join() for p in proc]

    return [x for i, x in sorted(res)]


if __name__ == '__main__':
    print(parmap(lambda i: i * 2, [1, 2, 3, 4, 6, 7, 8]))

Question 25

Multiprocessing and pickling is broken and limited unless you jump outside the standard library.

If you use a fork of multiprocessing called pathos.multiprocesssing, you can directly use classes and class methods in multiprocessing’s map functions. This is because dill is used instead of pickle or cPickle, and dill can serialize almost anything in python.

pathos.multiprocessing also provides an asynchronous map function… and it can map functions with multiple arguments (e.g. map(math.pow, [1,2,3], [4,5,6]))

See discussions: What can multiprocessing and dill do together?

and: http://matthewrocklin.com/blog/work/2013/12/05/Parallelism-and-Serialization

It even handles the code you wrote initially, without modification, and from the interpreter. Why do anything else that’s more fragile and specific to a single case?

>>> from pathos.multiprocessing import ProcessingPool as Pool
>>> class calculate(object):
...  def run(self):
...   def f(x):
...    return x*x
...   p = Pool()
...   return p.map(f, [1,2,3])
... 
>>> cl = calculate()
>>> print cl.run()
[1, 4, 9]

Get the code here: https://github.com/uqfoundation/pathos

And, just to show off a little more of what it can do:

>>> from pathos.multiprocessing import ProcessingPool as Pool
>>> 
>>> p = Pool(4)
>>> 
>>> def add(x,y):
...   return x+y
... 
>>> x = [0,1,2,3]
>>> y = [4,5,6,7]
>>> 
>>> p.map(add, x, y)
[4, 6, 8, 10]
>>> 
>>> class Test(object):
...   def plus(self, x, y): 
...     return x+y
... 
>>> t = Test()
>>> 
>>> p.map(Test.plus, [t]*4, x, y)
[4, 6, 8, 10]
>>> 
>>> res = p.amap(t.plus, x, y)
>>> res.get()
[4, 6, 8, 10]

Question 26

There is currently no solution to your problem, as far as I know: the function that you give to map() must be accessible through an import of your module. This is why robert’s code works: the function f() can be obtained by importing the following code:

def f(x):
    return x*x

class Calculate(object):
    def run(self):
        p = Pool()
        return p.map(f, [1,2,3])

if __name__ == '__main__':
    cl = Calculate()
    print cl.run()

I actually added a “main” section, because this follows the recommendations for the Windows platform (“Make sure that the main module can be safely imported by a new Python interpreter without causing unintended side effects”).

I also added an uppercase letter in front of Calculate, so as to follow PEP 8. :)

Question 27

The solution by mrule is correct but has a bug: if the child sends back a large amount of data, it can fill the pipe’s buffer, blocking on the child’s pipe.send(), while the parent is waiting for the child to exit on pipe.join(). The solution is to read the child’s data before join()ing the child. Furthermore the child should close the parent’s end of the pipe to prevent a deadlock. The code below fixes that. Also be aware that this parmap creates one process per element in X. A more advanced solution is to use multiprocessing.cpu_count() to divide X into a number of chunks, and then merge the results before returning. I leave that as an exercise to the reader so as not to spoil the conciseness of the nice answer by mrule. ;)

from multiprocessing import Process, Pipe
from itertools import izip

def spawn(f):
    def fun(ppipe, cpipe,x):
        ppipe.close()
        cpipe.send(f(x))
        cpipe.close()
    return fun

def parmap(f,X):
    pipe=[Pipe() for x in X]
    proc=[Process(target=spawn(f),args=(p,c,x)) for x,(p,c) in izip(X,pipe)]
    [p.start() for p in proc]
    ret = [p.recv() for (p,c) in pipe]
    [p.join() for p in proc]
    return ret

if __name__ == '__main__':
    print parmap(lambda x:x**x,range(1,5))

Question 28

I’ve also struggled with this. I had functions as data members of a class, as a simplified example:

from multiprocessing import Pool
import itertools
pool = Pool()
class Example(object):
    def __init__(self, my_add): 
        self.f = my_add  
    def add_lists(self, list1, list2):
        # Needed to do something like this (the following line won't work)
        return pool.map(self.f,list1,list2)

I needed to use the function self.f in a Pool.map() call from within the same class and self.f did not take a tuple as an argument. Since this function was embedded in a class, it was not clear to me how to write the type of wrapper other answers suggested.

I solved this problem by using a different wrapper that takes a tuple/list, where the first element is the function, and the remaining elements are the arguments to that function, called eval_func_tuple(f_args). Using this, the problematic line can be replaced by return pool.map(eval_func_tuple, itertools.izip(itertools.repeat(self.f), list1, list2)). Here is the full code:

File: util.py

def add(a, b): return a+b

def eval_func_tuple(f_args):
    """Takes a tuple of a function and args, evaluates and returns result"""
    return f_args[0](*f_args[1:])

File: main.py

from multiprocessing import Pool
import itertools
import util  

pool = Pool()
class Example(object):
    def __init__(self, my_add): 
        self.f = my_add  
    def add_lists(self, list1, list2):
        # The following line will now work
        return pool.map(util.eval_func_tuple, 
            itertools.izip(itertools.repeat(self.f), list1, list2)) 

if __name__ == '__main__':
    myExample = Example(util.add)
    list1 = [1, 2, 3]
    list2 = [10, 20, 30]
    print myExample.add_lists(list1, list2)

Running main.py will give [11, 22, 33]. Feel free to improve this, for example eval_func_tuple could also be modified to take keyword arguments.

On another note, in another answers, the function “parmap” can be made more efficient for the case of more Processes than number of CPUs available. I’m copying an edited version below. This is my first post and I wasn’t sure if I should directly edit the original answer. I also renamed some variables.

from multiprocessing import Process, Pipe  
from itertools import izip  

def spawn(f):  
    def fun(pipe,x):  
        pipe.send(f(x))  
        pipe.close()  
    return fun  

def parmap(f,X):  
    pipe=[Pipe() for x in X]  
    processes=[Process(target=spawn(f),args=(c,x)) for x,(p,c) in izip(X,pipe)]  
    numProcesses = len(processes)  
    processNum = 0  
    outputList = []  
    while processNum < numProcesses:  
        endProcessNum = min(processNum+multiprocessing.cpu_count(), numProcesses)  
        for proc in processes[processNum:endProcessNum]:  
            proc.start()  
        for proc in processes[processNum:endProcessNum]:  
            proc.join()  
        for proc,c in pipe[processNum:endProcessNum]:  
            outputList.append(proc.recv())  
        processNum = endProcessNum  
    return outputList    

if __name__ == '__main__':  
    print parmap(lambda x:x**x,range(1,5))

Question 29

I took klaus se’s and aganders3’s answer, and made a documented module that is more readable and holds in one file. You can just add it to your project. It even has an optional progress bar !

"""
The ``processes`` module provides some convenience functions
for using parallel processes in python.

Adapted from http://stackoverflow.com/a/16071616/287297

Example usage:

    print prll_map(lambda i: i * 2, [1, 2, 3, 4, 6, 7, 8], 32, verbose=True)

Comments:

"It spawns a predefined amount of workers and only iterates through the input list
 if there exists an idle worker. I also enabled the "daemon" mode for the workers so
 that KeyboardInterupt works as expected."

Pitfalls: all the stdouts are sent back to the parent stdout, intertwined.

Alternatively, use this fork of multiprocessing: 
https://github.com/uqfoundation/multiprocess
"""

# Modules #
import multiprocessing
from tqdm import tqdm

################################################################################
def apply_function(func_to_apply, queue_in, queue_out):
    while not queue_in.empty():
        num, obj = queue_in.get()
        queue_out.put((num, func_to_apply(obj)))

################################################################################
def prll_map(func_to_apply, items, cpus=None, verbose=False):
    # Number of processes to use #
    if cpus is None: cpus = min(multiprocessing.cpu_count(), 32)
    # Create queues #
    q_in  = multiprocessing.Queue()
    q_out = multiprocessing.Queue()
    # Process list #
    new_proc  = lambda t,a: multiprocessing.Process(target=t, args=a)
    processes = [new_proc(apply_function, (func_to_apply, q_in, q_out)) for x in range(cpus)]
    # Put all the items (objects) in the queue #
    sent = [q_in.put((i, x)) for i, x in enumerate(items)]
    # Start them all #
    for proc in processes:
        proc.daemon = True
        proc.start()
    # Display progress bar or not #
    if verbose:
        results = [q_out.get() for x in tqdm(range(len(sent)))]
    else:
        results = [q_out.get() for x in range(len(sent))]
    # Wait for them to finish #
    for proc in processes: proc.join()
    # Return results #
    return [x for i, x in sorted(results)]

################################################################################
def test():
    def slow_square(x):
        import time
        time.sleep(2)
        return x**2
    objs    = range(20)
    squares = prll_map(slow_square, objs, 4, verbose=True)
    print "Result: %s" % squares

EDIT: Added @alexander-mcfarlane suggestion and a test function

Question 30

I know this was asked over 6 years ago now, but just wanted to add my solution, as some of the suggestions above seem horribly complicated, but my solution was actually very simple.

All I had to do was wrap the pool.map() call to a helper function. Passing the class object along with args for the method as a tuple, which looked a bit like this.

def run_in_parallel(args):
    return args[0].method(args[1])

myclass = MyClass()
method_args = [1,2,3,4,5,6]
args_map = [ (myclass, arg) for arg in method_args ]
pool = Pool()
pool.map(run_in_parallel, args_map)

Question 31

Functions defined in classes (even within functions within classes) don’t really pickle. However, this works:

def f(x):
    return x*x

class calculate(object):
    def run(self):
        p = Pool()
    return p.map(f, [1,2,3])

cl = calculate()
print cl.run()

Question 32

I know that this question was asked 8 years and 10 months ago but I want to present you my solution:

from multiprocessing import Pool

class Test:

    def __init__(self):
        self.main()

    @staticmethod
    def methodForMultiprocessing(x):
        print(x*x)

    def main(self):
        if __name__ == "__main__":
            p = Pool()
            p.map(Test.methodForMultiprocessing, list(range(1, 11)))
            p.close()

TestObject = Test()

You just need to make your class function into a static method. But it’s also possible with a class method:

from multiprocessing import Pool

class Test:

    def __init__(self):
        self.main()

    @classmethod
    def methodForMultiprocessing(cls, x):
        print(x*x)

    def main(self):
        if __name__ == "__main__":
            p = Pool()
            p.map(Test.methodForMultiprocessing, list(range(1, 11)))
            p.close()

TestObject = Test()

Tested in Python 3.7.3

Question 33

I modified klaus se’s method because while it was working for me with small lists, it would hang when the number of items was ~1000 or greater. Instead of pushing the jobs one at a time with the None stop condition, I load up the input queue all at once and just let the processes munch on it until it’s empty.

from multiprocessing import cpu_count, Queue, Process

def apply_func(f, q_in, q_out):
    while not q_in.empty():
        i, x = q_in.get()
        q_out.put((i, f(x)))

# map a function using a pool of processes
def parmap(f, X, nprocs = cpu_count()):
    q_in, q_out   = Queue(), Queue()
    proc = [Process(target=apply_func, args=(f, q_in, q_out)) for _ in range(nprocs)]
    sent = [q_in.put((i, x)) for i, x in enumerate(X)]
    [p.start() for p in proc]
    res = [q_out.get() for _ in sent]
    [p.join() for p in proc]

    return [x for i,x in sorted(res)]

Edit: unfortunately now I am running into this error on my system: Multiprocessing Queue maxsize limit is 32767, hopefully the workarounds there will help.

Question 34

You can run your code without any issues if you somehow manually ignore the Pool object from the list of objects in the class because it is not pickleable as the error says. You can do this with the __getstate__ function (look here too) as follow. The Pool object will try to find the __getstate__ and __setstate__ functions and execute them if it finds it when you run map, map_async etc:

class calculate(object):
    def __init__(self):
        self.p = Pool()
    def __getstate__(self):
        self_dict = self.__dict__.copy()
        del self_dict['p']
        return self_dict
    def __setstate__(self, state):
        self.__dict__.update(state)

    def f(self, x):
        return x*x
    def run(self):
        return self.p.map(self.f, [1,2,3])

Then do:

cl = calculate()
cl.run()

will give you the output:

[1, 4, 9]

I’ve tested the above code in Python 3.x and it works.

Question 35

I’m not sure if this approach has been taken but a work around i’m using is:

from multiprocessing import Pool

t = None

def run(n):
    return t.f(n)

class Test(object):
    def __init__(self, number):
        self.number = number

    def f(self, x):
        print x * self.number

    def pool(self):
        pool = Pool(2)
        pool.map(run, range(10))

if __name__ == '__main__':
    t = Test(9)
    t.pool()
    pool = Pool(2)
    pool.map(run, range(10))

Output should be:

Question 36

class Calculate(object):
  # Your instance method to be executed
  def f(self, x, y):
    return x*y

if __name__ == '__main__':
  inp_list = [1,2,3]
  y = 2
  cal_obj = Calculate()
  pool = Pool(2)
  results = pool.map(lambda x: cal_obj.f(x, y), inp_list)

There is a possibility that you would want to apply this function for each different instance of the class. Then here is the solution for that also

class Calculate(object):
  # Your instance method to be executed
  def __init__(self, x):
    self.x = x

  def f(self, y):
    return self.x*y

if __name__ == '__main__':
  inp_list = [Calculate(i) for i in range(3)]
  y = 2
  pool = Pool(2)
  results = pool.map(lambda x: x.f(y), inp_list)

Question 37

Here is my solution, which I think is a bit less hackish than most others here. It is similar to nightowl’s answer.

someclasses = [MyClass(), MyClass(), MyClass()]

def method_caller(some_object, some_method='the method'):
    return getattr(some_object, some_method)()

othermethod = partial(method_caller, some_method='othermethod')

with Pool(6) as pool:
    result = pool.map(othermethod, someclasses)

Question 38

From http://www.rueckstiess.net/research/snippets/show/ca1d7d90 and http://qingkaikong.blogspot.com/2016/12/python-parallel-method-in-class.html

We can make an external function and seed it with the class self object:

from joblib import Parallel, delayed
def unwrap_self(arg, **kwarg):
    return square_class.square_int(*arg, **kwarg)

class square_class:
    def square_int(self, i):
        return i * i

    def run(self, num):
        results = []
        results = Parallel(n_jobs= -1, backend="threading")\
            (delayed(unwrap_self)(i) for i in zip([self]*len(num), num))
        print(results)

OR without joblib:

from multiprocessing import Pool
import time

def unwrap_self_f(arg, **kwarg):
    return C.f(*arg, **kwarg)

class C:
    def f(self, name):
        print 'hello %s,'%name
        time.sleep(5)
        print 'nice to meet you.'

    def run(self):
        pool = Pool(processes=2)
        names = ('frank', 'justin', 'osi', 'thomas')
        pool.map(unwrap_self_f, zip([self]*len(names), names))

if __name__ == '__main__':
    c = C()
    c.run()

Question 39

This may not be a very good solution but in my case, I solve it like this.

from multiprocessing import Pool

def foo1(data):
    self = data.get('slf')
    lst = data.get('lst')
    return sum(lst) + self.foo2()

class Foo(object):
    def __init__(self, a, b):
        self.a = a
        self.b = b

    def foo2(self):
        return self.a**self.b   

    def foo(self):
        p = Pool(5)
        lst = [1, 2, 3]
        result = p.map(foo1, (dict(slf=self, lst=lst),))
        return result

if __name__ == '__main__':
    print(Foo(2, 4).foo())

I had to pass self to my function as I have to access attributes and functions of my class through that function. This is working for me. Corrections and suggestions are always welcome.

Question 40

Here is a boilerplate I wrote for using multiprocessing Pool in python3, specifically python3.7.7 was used to run the tests. I got my fastest runs using imap_unordered. Just plug in your scenario and try it out. You can use timeit or just time.time() to figure out which works best for you.

import multiprocessing
import time

NUMBER_OF_PROCESSES = multiprocessing.cpu_count()
MP_FUNCTION = 'starmap'  # 'imap_unordered' or 'starmap' or 'apply_async'

def process_chunk(a_chunk):
    print(f"processig mp chunk {a_chunk}")
    return a_chunk


map_jobs = [1, 2, 3, 4]

result_sum = 0

s = time.time()
if MP_FUNCTION == 'imap_unordered':
    pool = multiprocessing.Pool(processes=NUMBER_OF_PROCESSES)
    for i in pool.imap_unordered(process_chunk, map_jobs):
        result_sum += i
elif MP_FUNCTION == 'starmap':
    pool = multiprocessing.Pool(processes=NUMBER_OF_PROCESSES)
    try:
        map_jobs = [(i, ) for i in map_jobs]
        result_sum = pool.starmap(process_chunk, map_jobs)
        result_sum = sum(result_sum)
    finally:
        pool.close()
        pool.join()
elif MP_FUNCTION == 'apply_async':
    with multiprocessing.Pool(processes=NUMBER_OF_PROCESSES) as pool:
        result_sum = [pool.apply_async(process_chunk, [i, ]).get() for i in map_jobs]
    result_sum = sum(result_sum)
print(f"result_sum is {result_sum}, took {time.time() - s}s")

In the above scenario imap_unordered actually seems to perform the worst for me. Try out your case and benchmark it on the machine you plan to run it on. Also read up on Process Pools. Cheers!

Question 41

I am sorry that I can’t reproduce the error with a simpler example, and my code is too complicated to post. If I run the program in IPython shell instead of the regular Python, things work out well.

I looked up some previous notes on this problem. They were all caused by using pool to call function defined within a class function. But this is not the case for me.

Exception in thread Thread-3:
Traceback (most recent call last):
  File "/usr/lib64/python2.7/threading.py", line 552, in __bootstrap_inner
    self.run()
  File "/usr/lib64/python2.7/threading.py", line 505, in run
    self.__target(*self.__args, **self.__kwargs)
  File "/usr/lib64/python2.7/multiprocessing/pool.py", line 313, in _handle_tasks
    put(task)
PicklingError: Can't pickle <type 'function'>: attribute lookup __builtin__.function failed

I would appreciate any help.

Update: The function I pickle is defined at the top level of the module. Though it calls a function that contains a nested function. i.e, f() calls g() calls h() which has a nested function i(), and I am calling pool.apply_async(f). f(), g(), h() are all defined at the top level. I tried simpler example with this pattern and it works though.

Question 42

Here is a list of what can be pickled. In particular, functions are only picklable if they are defined at the top-level of a module.

This piece of code:

import multiprocessing as mp

class Foo():
    @staticmethod
    def work(self):
        pass

if __name__ == '__main__':   
    pool = mp.Pool()
    foo = Foo()
    pool.apply_async(foo.work)
    pool.close()
    pool.join()

yields an error almost identical to the one you posted:

Exception in thread Thread-2:
Traceback (most recent call last):
  File "/usr/lib/python2.7/threading.py", line 552, in __bootstrap_inner
    self.run()
  File "/usr/lib/python2.7/threading.py", line 505, in run
    self.__target(*self.__args, **self.__kwargs)
  File "/usr/lib/python2.7/multiprocessing/pool.py", line 315, in _handle_tasks
    put(task)
PicklingError: Can't pickle <type 'function'>: attribute lookup __builtin__.function failed

The problem is that the pool methods all use a mp.SimpleQueue to pass tasks to the worker processes. Everything that goes through the mp.SimpleQueue must be pickable, and foo.work is not picklable since it is not defined at the top level of the module.

It can be fixed by defining a function at the top level, which calls foo.work():

def work(foo):
    foo.work()

pool.apply_async(work,args=(foo,))

Notice that foo is pickable, since Foo is defined at the top level and foo.__dict__ is picklable.

Question 43

I’d use pathos.multiprocesssing, instead of multiprocessing. pathos.multiprocessing is a fork of multiprocessing that uses dill. dill can serialize almost anything in python, so you are able to send a lot more around in parallel. The pathos fork also has the ability to work directly with multiple argument functions, as you need for class methods.

>>> from pathos.multiprocessing import ProcessingPool as Pool
>>> p = Pool(4)
>>> class Test(object):
...   def plus(self, x, y): 
...     return x+y
... 
>>> t = Test()
>>> p.map(t.plus, x, y)
[4, 6, 8, 10]
>>> 
>>> class Foo(object):
...   @staticmethod
...   def work(self, x):
...     return x+1
... 
>>> f = Foo()
>>> p.apipe(f.work, f, 100)
<processing.pool.ApplyResult object at 0x10504f8d0>
>>> res = _
>>> res.get()
101

Get pathos (and if you like, dill) here: https://github.com/uqfoundation

Question 44

As others have said multiprocessing can only transfer Python objects to worker processes which can be pickled. If you cannot reorganize your code as described by unutbu, you can use dills extended pickling/unpickling capabilities for transferring data (especially code data) as I show below.

This solution requires only the installation of dill and no other libraries as pathos:

import os
from multiprocessing import Pool

import dill


def run_dill_encoded(payload):
    fun, args = dill.loads(payload)
    return fun(*args)


def apply_async(pool, fun, args):
    payload = dill.dumps((fun, args))
    return pool.apply_async(run_dill_encoded, (payload,))


if __name__ == "__main__":

    pool = Pool(processes=5)

    # asyn execution of lambda
    jobs = []
    for i in range(10):
        job = apply_async(pool, lambda a, b: (a, b, a * b), (i, i + 1))
        jobs.append(job)

    for job in jobs:
        print job.get()
    print

    # async execution of static method

    class O(object):

        @staticmethod
        def calc():
            return os.getpid()

    jobs = []
    for i in range(10):
        job = apply_async(pool, O.calc, ())
        jobs.append(job)

    for job in jobs:
        print job.get()

Question 45

I have found that I can also generate exactly that error output on a perfectly working piece of code by attempting to use the profiler on it.

Note that this was on Windows (where the forking is a bit less elegant).

I was running:

python -m profile -o output.pstats <script>

And found that removing the profiling removed the error and placing the profiling restored it. Was driving me batty too because I knew the code used to work. I was checking to see if something had updated pool.py… then had a sinking feeling and eliminated the profiling and that was it.

Posting here for the archives in case anybody else runs into it.

Question 46

When this problem comes up with multiprocessing a simple solution is to switch from Pool to ThreadPool. This can be done with no change of code other than the import-

from multiprocessing.pool import ThreadPool as Pool

This works because ThreadPool shares memory with the main thread, rather than creating a new process- this means that pickling is not required.

The downside to this method is that python isn’t the greatest language with handling threads- it uses something called the Global Interpreter Lock to stay thread safe, which can slow down some use cases here. However, if you’re primarily interacting with other systems (running HTTP commands, talking with a database, writing to filesystems) then your code is likely not bound by CPU and won’t take much of a hit. In fact I’ve found when writing HTTP/HTTPS benchmarks that the threaded model used here has less overhead and delays, as the overhead from creating new processes is much higher than the overhead for creating new threads.

So if you’re processing a ton of stuff in python userspace this might not be the best method.

Question 47

This solution requires only the installation of dill and no other libraries as pathos

def apply_packed_function_for_map((dumped_function, item, args, kwargs),):
    """
    Unpack dumped function as target function and call it with arguments.

    :param (dumped_function, item, args, kwargs):
        a tuple of dumped function and its arguments
    :return:
        result of target function
    """
    target_function = dill.loads(dumped_function)
    res = target_function(item, *args, **kwargs)
    return res


def pack_function_for_map(target_function, items, *args, **kwargs):
    """
    Pack function and arguments to object that can be sent from one
    multiprocessing.Process to another. The main problem is:
        «multiprocessing.Pool.map*» or «apply*»
        cannot use class methods or closures.
    It solves this problem with «dill».
    It works with target function as argument, dumps it («with dill»)
    and returns dumped function with arguments of target function.
    For more performance we dump only target function itself
    and don't dump its arguments.
    How to use (pseudo-code):

        ~>>> import multiprocessing
        ~>>> images = [...]
        ~>>> pool = multiprocessing.Pool(100500)
        ~>>> features = pool.map(
        ~...     *pack_function_for_map(
        ~...         super(Extractor, self).extract_features,
        ~...         images,
        ~...         type='png'
        ~...         **options,
        ~...     )
        ~... )
        ~>>>

    :param target_function:
        function, that you want to execute like  target_function(item, *args, **kwargs).
    :param items:
        list of items for map
    :param args:
        positional arguments for target_function(item, *args, **kwargs)
    :param kwargs:
        named arguments for target_function(item, *args, **kwargs)
    :return: tuple(function_wrapper, dumped_items)
        It returs a tuple with
            * function wrapper, that unpack and call target function;
            * list of packed target function and its' arguments.
    """
    dumped_function = dill.dumps(target_function)
    dumped_items = [(dumped_function, item, args, kwargs) for item in items]
    return apply_packed_function_for_map, dumped_items

It also works for numpy arrays.

Question 48

Can't pickle <type 'function'>: attribute lookup __builtin__.function failed

This error will also come if you have any inbuilt function inside the model object that was passed to the async job.

So make sure to check the model objects that are passed doesn’t have inbuilt functions. (In our case we were using FieldTracker() function of django-model-utils inside the model to track a certain field). Here is the link to relevant GitHub issue.

Question 49

Building on @rocksportrocker solution, It would make sense to dill when sending and RECVing the results.

import dill
import itertools
def run_dill_encoded(payload):
    fun, args = dill.loads(payload)
    res = fun(*args)
    res = dill.dumps(res)
    return res

def dill_map_async(pool, fun, args_list,
                   as_tuple=True,
                   **kw):
    if as_tuple:
        args_list = ((x,) for x in args_list)

    it = itertools.izip(
        itertools.cycle([fun]),
        args_list)
    it = itertools.imap(dill.dumps, it)
    return pool.map_async(run_dill_encoded, it, **kw)

if __name__ == '__main__':
    import multiprocessing as mp
    import sys,os
    p = mp.Pool(4)
    res = dill_map_async(p, lambda x:[sys.stdout.write('%s\n'%os.getpid()),x][-1],
                  [lambda x:x+1]*10,)
    res = res.get(timeout=100)
    res = map(dill.loads,res)
    print(res)

Question 50

I’m using python3.3 and I’m having a cryptic error when trying to pickle a simple dictionary.

Here is the code:

import os
import pickle
from pickle import *
os.chdir('c:/Python26/progfiles/')

def storvars(vdict):      
    f = open('varstor.txt','w')
    pickle.dump(vdict,f,)
    f.close()
    return

mydict = {'name':'john','gender':'male','age':'45'}
storvars(mydict)

and I get:

Traceback (most recent call last):
  File "C:/Python26/test18.py", line 31, in <module>
    storvars(mydict)
  File "C:/Python26/test18.py", line 14, in storvars
    pickle.dump(vdict,f,)
TypeError: must be str, not bytes

Question 51

The output file needs to be opened in binary mode:

f = open('varstor.txt','w')

needs to be:

f = open('varstor.txt','wb')

Question 52

Just had same issue. In Python 3, Binary modes ‘wb’, ‘rb’ must be specified whereas in Python 2x, they are not needed. When you follow tutorials that are based on Python 2x, that’s why you are here.

import pickle

class MyUser(object):
    def __init__(self,name):
        self.name = name

user = MyUser('Peter')

print("Before serialization: ")
print(user.name)
print("------------")
serialized = pickle.dumps(user)
filename = 'serialized.native'

with open(filename,'wb') as file_object:
    file_object.write(serialized)

with open(filename,'rb') as file_object:
    raw_data = file_object.read()

deserialized = pickle.loads(raw_data)


print("Loading from serialized file: ")
user2 = deserialized
print(user2.name)
print("------------")

Question 53

I’ve created an object like this:

company1.name = 'banana' 
company1.value = 40

I would like to save this object. How can I do that?

Question 54

You could use the pickle module in the standard library. Here’s an elementary application of it to your example:

import pickle

class Company(object):
    def __init__(self, name, value):
        self.name = name
        self.value = value

with open('company_data.pkl', 'wb') as output:
    company1 = Company('banana', 40)
    pickle.dump(company1, output, pickle.HIGHEST_PROTOCOL)

    company2 = Company('spam', 42)
    pickle.dump(company2, output, pickle.HIGHEST_PROTOCOL)

del company1
del company2

with open('company_data.pkl', 'rb') as input:
    company1 = pickle.load(input)
    print(company1.name)  # -> banana
    print(company1.value)  # -> 40

    company2 = pickle.load(input)
    print(company2.name) # -> spam
    print(company2.value)  # -> 42

You could also define your own simple utility like the following which opens a file and writes a single object to it:

def save_object(obj, filename):
    with open(filename, 'wb') as output:  # Overwrites any existing file.
        pickle.dump(obj, output, pickle.HIGHEST_PROTOCOL)

# sample usage
save_object(company1, 'company1.pkl')

Update

Since this is such a popular answer, I’d like touch on a few slightly advanced usage topics.

`cPickle` (or `_pickle`) vs `pickle`

It’s almost always preferable to actually use the cPickle module rather than pickle because the former is written in C and is much faster. There are some subtle differences between them, but in most situations they’re equivalent and the C version will provide greatly superior performance. Switching to it couldn’t be easier, just change the import statement to this:

import cPickle as pickle

In Python 3, cPickle was renamed _pickle, but doing this is no longer necessary since the pickle module now does it automatically—see What difference between pickle and _pickle in python 3?.

The rundown is you could use something like the following to ensure that your code will always use the C version when it’s available in both Python 2 and 3:

try:
    import cPickle as pickle
except ModuleNotFoundError:
    import pickle

Data stream formats (protocols)

pickle can read and write files in several different, Python-specific, formats, called protocols as described in the documentation, “Protocol version 0” is ASCII and therefore “human-readable”. Versions > 0 are binary and the highest one available depends on what version of Python is being used. The default also depends on Python version. In Python 2 the default was Protocol version 0, but in Python 3.8.1, it’s Protocol version 4. In Python 3.x the module had a pickle.DEFAULT_PROTOCOL added to it, but that doesn’t exist in Python 2.

Fortunately there’s shorthand for writing pickle.HIGHEST_PROTOCOL in every call (assuming that’s what you want, and you usually do), just use the literal number -1 — similar to referencing the last element of a sequence via a negative index. So, instead of writing:

pickle.dump(obj, output, pickle.HIGHEST_PROTOCOL)

You can just write:

pickle.dump(obj, output, -1)

Either way, you’d only have specify the protocol once if you created a Pickler object for use in multiple pickle operations:

pickler = pickle.Pickler(output, -1)
pickler.dump(obj1)
pickler.dump(obj2)
   etc...

Note: If you’re in an environment running different versions of Python, then you’ll probably want to explicitly use (i.e. hardcode) a specific protocol number that all of them can read (later versions can generally read files produced by earlier ones).

Multiple Objects

While a pickle file can contain any number of pickled objects, as shown in the above samples, when there’s an unknown number of them, it’s often easier to store them all in some sort of variably-sized container, like a list, tuple, or dict and write them all to the file in a single call:

tech_companies = [
    Company('Apple', 114.18), Company('Google', 908.60), Company('Microsoft', 69.18)
]
save_object(tech_companies, 'tech_companies.pkl')

and restore the list and everything in it later with:

with open('tech_companies.pkl', 'rb') as input:
    tech_companies = pickle.load(input)

The major advantage is you don’t need to know how many object instances are saved in order to load them back later (although doing so without that information is possible, it requires some slightly specialized code). See the answers to the related question Saving and loading multiple objects in pickle file? for details on different ways to do this. Personally I like @Lutz Prechelt’s answer the best. Here’s it adapted to the examples here:

class Company:
    def __init__(self, name, value):
        self.name = name
        self.value = value

def pickled_items(filename):
    """ Unpickle a file of pickled data. """
    with open(filename, "rb") as f:
        while True:
            try:
                yield pickle.load(f)
            except EOFError:
                break

print('Companies in pickle file:')
for company in pickled_items('company_data.pkl'):
    print('  name: {}, value: {}'.format(company.name, company.value))

Question 55

I think it’s a pretty strong assumption to assume that the object is a class. What if it’s not a class? There’s also the assumption that the object was not defined in the interpreter. What if it was defined in the interpreter? Also, what if the attributes were added dynamically? When some python objects have attributes added to their __dict__ after creation, pickle doesn’t respect the addition of those attributes (i.e. it ‘forgets’ they were added — because pickle serializes by reference to the object definition).

In all these cases, pickle and cPickle can fail you horribly.

If you are looking to save an object (arbitrarily created), where you have attributes (either added in the object definition, or afterward)… your best bet is to use dill, which can serialize almost anything in python.

We start with a class…

Python 2.7.8 (default, Jul 13 2014, 02:29:54) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import pickle
>>> class Company:
...     pass
... 
>>> company1 = Company()
>>> company1.name = 'banana'
>>> company1.value = 40
>>> with open('company.pkl', 'wb') as f:
...     pickle.dump(company1, f, pickle.HIGHEST_PROTOCOL)
... 
>>>

Now shut down, and restart…

Python 2.7.8 (default, Jul 13 2014, 02:29:54) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import pickle
>>> with open('company.pkl', 'rb') as f:
...     company1 = pickle.load(f)
... 
Traceback (most recent call last):
  File "<stdin>", line 2, in <module>
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 1378, in load
    return Unpickler(file).load()
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 858, in load
dispatch[key](self)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 1090, in load_global
    klass = self.find_class(module, name)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/pickle.py", line 1126, in find_class
    klass = getattr(mod, name)
AttributeError: 'module' object has no attribute 'Company'
>>>

Oops… pickle can’t handle it. Let’s try dill. We’ll throw in another object type (a lambda) for good measure.

Python 2.7.8 (default, Jul 13 2014, 02:29:54) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import dill       
>>> class Company:
...     pass
... 
>>> company1 = Company()
>>> company1.name = 'banana'
>>> company1.value = 40
>>> 
>>> company2 = lambda x:x
>>> company2.name = 'rhubarb'
>>> company2.value = 42
>>> 
>>> with open('company_dill.pkl', 'wb') as f:
...     dill.dump(company1, f)
...     dill.dump(company2, f)
... 
>>>

And now read the file.

Python 2.7.8 (default, Jul 13 2014, 02:29:54) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import dill
>>> with open('company_dill.pkl', 'rb') as f:
...     company1 = dill.load(f)
...     company2 = dill.load(f)
... 
>>> company1 
<__main__.Company instance at 0x107909128>
>>> company1.name
'banana'
>>> company1.value
40
>>> company2.name
'rhubarb'
>>> company2.value
42
>>>

It works. The reason pickle fails, and dill doesn’t, is that dill treats __main__ like a module (for the most part), and also can pickle class definitions instead of pickling by reference (like pickle does). The reason dill can pickle a lambda is that it gives it a name… then pickling magic can happen.

Actually, there’s an easier way to save all these objects, especially if you have a lot of objects you’ve created. Just dump the whole python session, and come back to it later.

Python 2.7.8 (default, Jul 13 2014, 02:29:54) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import dill
>>> class Company:
...     pass
... 
>>> company1 = Company()
>>> company1.name = 'banana'
>>> company1.value = 40
>>> 
>>> company2 = lambda x:x
>>> company2.name = 'rhubarb'
>>> company2.value = 42
>>> 
>>> dill.dump_session('dill.pkl')
>>>

Now shut down your computer, go enjoy an espresso or whatever, and come back later…

Python 2.7.8 (default, Jul 13 2014, 02:29:54) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import dill
>>> dill.load_session('dill.pkl')
>>> company1.name
'banana'
>>> company1.value
40
>>> company2.name
'rhubarb'
>>> company2.value
42
>>> company2
<function <lambda> at 0x1065f2938>

The only major drawback is that dill is not part of the python standard library. So if you can’t install a python package on your server, then you can’t use it.

However, if you are able to install python packages on your system, you can get the latest dill with git+https://github.com/uqfoundation/dill.git@master#egg=dill. And you can get the latest released version with pip install dill.

Question 56

You can use anycache to do the job for you. It considers all the details:

It uses dill as backend, which extends the python pickle module to handle lambda and all the nice python features.
It stores different objects to different files and reloads them properly.
Limits cache size
Allows cache clearing
Allows sharing of objects between multiple runs
Allows respect of input files which influence the result

Assuming you have a function myfunc which creates the instance:

from anycache import anycache

class Company(object):
    def __init__(self, name, value):
        self.name = name
        self.value = value

@anycache(cachedir='/path/to/your/cache')    
def myfunc(name, value)
    return Company(name, value)

Anycache calls myfunc at the first time and pickles the result to a file in cachedir using an unique identifier (depending on the function name and its arguments) as filename. On any consecutive run, the pickled object is loaded. If the cachedir is preserved between python runs, the pickled object is taken from the previous python run.

For any further details see the documentation

Question 57

Quick example using company1 from your question, with python3.

import pickle

# Save the file
pickle.dump(company1, file = open("company1.pickle", "wb"))

# Reload the file
company1_reloaded = pickle.load(open("company1.pickle", "rb"))

However, as this answer noted, pickle often fails. So you should really use dill.

import dill

# Save the file
dill.dump(company1, file = open("company1.pickle", "wb"))

# Reload the file
company1_reloaded = dill.load(open("company1.pickle", "rb"))

问题：存储Python字典

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

Console output

Contents of config.ini

回答 7

回答 8

问题：将类实例序列化为JSON

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

问题：多重处理：如何在类中定义的函数上使用Pool.map？

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

回答 10

回答 11

回答 12

回答 13

回答 14

回答 15

回答 16

回答 17

问题：Python多处理PicklingError：无法腌制

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

问题：使用pickle.dump-TypeError：必须为str，而不是字节

回答 0

回答 1

问题：保存对象（数据持久性）

回答 0

更新资料

cPickle（或_pickle）与pickle

数据流格式（协议）

多个物件

Update

cPickle (or _pickle) vs pickle

Data stream formats (protocols)

Multiple Objects

回答 1

回答 2

回答 3

问题：如何使用泡菜保存字典？

回答 0

回答 1

替代方式

替代格式

Alternative way

Alternative Formats

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

`cPickle`（或`_pickle`）与`pickle`

`cPickle` (or `_pickle`) vs `pickle`