python3 TypeError：“字节”对象不可调用

Question

My code:

 for i in range(data.num_nodes):
        if embed[i]:
            # print embed[i]
            tmp = np.sum(embed[i], axis=0) / len(embed[i])
            file.write(' '.join(map(str.encode("utf-8"), tmp)) + '\n')
        else:
            file.write('\n')

However, when I run code, I get:

file.write(' '.join(map(str.encode("utf-8"), tmp)) + '\n')
`TypeError: 'bytes' object is not callable`

When I change the code into this:

  for i in range(data.num_nodes):
        if embed[i]:
            # print embed[i]
            tmp = np.sum(embed[i], axis=0) / len(embed[i])
            file.write(' '.join(map(str, tmp)) + '\n')
        else:
            file.write('\n')

I get this error:

TypeError: a bytes-like object is required, not 'str'

Answer 1

Do:

file.write(' '.join(map(str.encode, tmp)) + '\n')

Instead of:

file.write(' '.join(map(str.encode("utf-8"), tmp)) + '\n')

Because str.encode needs a argument of a string and this works because default it's already utf-8 encoding

Answer 2

map expects a function object as its first argument, and with str.encode("utf-8") you actually call str.encode with 'utf-8' as the first argument and literally encode the string 'utf-8' to bytes, so when map calls the function in its first argument, it would fail as it is actually a string.

You should use functools.partial to pass str.encode to map str.encode as a function object with the encoding parameter pre-filled with your desired encoding:

from functools import partial
file.write(' '.join(map(partial(str.encode, encoding='utf-8'), tmp)) + '\n')

But since the encoding parameter for str.encode has a default value of 'utf-8' , you can simply make map use the defaults of str.encode by passing str.encode to map directly:

file.write(' '.join(map(str.encode, tmp)) + '\n')

But since what you really want to do is to convert the entire string you're going to pass to file.write as bytes, including both ' ' and '\\n' , which are strings rather than bytes, you should encode the entire string after the joining of the substrings and the concatenation with '\\n' :

file.write((' '.join(tmp) + '\n').encode())

And since your tmp is not a list of strings, but rather a list of numpy.float32 objects, you should map them to strings first before joining:

file.write((' '.join(map(str, tmp)) + '\n').encode())