您可以捕获该异常并从那里返回您想要的任何内容。

open(target, 'a').close()
scores = {};
try:
    with open(target, "rb") as file:
        unpickler = pickle.Unpickler(file);
        scores = unpickler.load();
        if not isinstance(scores, dict):
            scores = {};
except EOFError:
    return {}

我会先检查文件是否为空:

import os

scores = {} # scores is an empty dict already

if os.path.getsize(target) > 0:      
    with open(target, "rb") as f:
        unpickler = pickle.Unpickler(f)
        # if file is not empty scores will be equal
        # to the value unpickled
        scores = unpickler.load()

同样，open(target， 'a').close()在你的代码中什么也不做，你不需要使用;。

如你所见，这实际上是一个自然错误。

从Unpickler对象中读取数据的典型构造如下所示。

try:
    data = unpickler.load()
except EOFError:
    data = list()  # or whatever you want

简单地引发EOFError，因为它正在读取一个空文件，它只是意味着文件结束。

if path.exists(Score_file):
      try : 
         with open(Score_file , "rb") as prev_Scr:

            return Unpickler(prev_Scr).load()

    except EOFError : 

        return dict() 

这里的大多数答案都涉及了如何管理EOFError异常，如果您不确定pickle对象是否为空，这非常方便。

然而，如果您惊讶于pickle文件是空的，这可能是因为您通过'wb'或其他可能覆盖该文件的模式打开了文件名。

例如:

filename = 'cd.pkl'
with open(filename, 'wb') as f:
    classification_dict = pickle.load(f)

这将覆盖pickle文件。你可能在使用之前错误地这样做了:

...
open(filename, 'rb') as f:

然后得到EOFError，因为前面的代码块覆盖了cd.pkl文件。

当在Jupyter或在控制台(Spyder)中工作时，我通常在读写代码上编写包装器，然后调用包装器。这避免了常见的读写错误，如果要多次读取同一个文件，还节省了一些时间

pickle文件很可能是空的。

如果要复制和粘贴代码，那么覆盖pickle文件非常容易。

例如，下面写一个pickle文件:

pickle.dump(df,open('df.p','wb'))

如果你复制这段代码重新打开它，但忘记把'wb'改为'rb'，那么你会覆盖文件:

df=pickle.load(open('df.p','wb'))

正确的语法是

df=pickle.load(open('df.p','rb'))

注意，打开文件的模式是“a”或其他有字母的“a”也会因为覆盖而出错。

pointer = open('makeaafile.txt', 'ab+')
tes = pickle.load(pointer, encoding='utf-8')

我遇到过这个错误很多次，它总是发生，因为写入文件后，我没有关闭它。如果我们不关闭文件，内容保留在缓冲区中，文件保持为空。 要将内容保存到文件中，要么关闭文件，要么超出file_object的作用域。

这就是为什么在加载时它会给出run out of input错误因为文件是空的。所以你有两个选择:

file_object.close () File_object.flush():如果你不想在程序之间关闭文件，你可以使用flush()函数，因为它会强制将内容从缓冲区移动到文件中。

temp_model = os.path.join(models_dir, train_type + '_' + part + '_' + str(pc))
# print(type(temp_model)) # <class 'str'>
filehandler = open(temp_model, "rb")
# print(type(filehandler)) # <class '_io.BufferedReader'>
try:
    pdm_temp = pickle.load(filehandler)
except UnicodeDecodeError:
    pdm_temp = pickle.load(filehandler, fix_imports=True, encoding="latin1")

也有同样的问题。原来，当我写入pickle文件时，我没有使用file.close()。插入这一行，错误就不再存在了。

from os.path import getsize as size
from pickle import *
if size(target)>0:
    with open(target,'rb') as f:
        scores={i:j for i,j in enumerate(load(f))}
else: scores={}

#line 1. we importing Function 'getsize' from Library 'OS' sublibrary 'path' and we rename it with command 'as' for shorter style of writing. Important is hier that we loading only one single Func that we need and not whole Library! line 2. Same Idea, but when we dont know wich modul we will use in code at the begining, we can import all library using a command '*'. line 3. Conditional Statement... if size of your file >0 ( means obj is not an empty). 'target' is variable that schould be a bit earlier predefined. just an Example : target=(r'd:\dir1\dir.2..\YourDataFile.bin') Line 4. 'With open(target) as file:' an open construction for any file, u dont need then to use file.close(). it helps to avoid some typical Errors such as "Run out of input" or Permissions rights. 'rb' mod means 'rea binary' that u can only read(load) the data from your binary file but u cant modify/rewrite it. Line5. List comprehension method in applying to a Dictionary.. line 6. Case your datafile is empty, it will not raise an any Error msg, but return just an empty dictionary.

当pickle文件为空(0字节)时出现此错误。首先需要检查pickle文件的大小。这就是我的情况。希望这能有所帮助!