TypeError: can't pickle generator objects

Question

I am trying to write some result on to pickle file as below:

raw_X = (self.token_ques(text) for text in training_data)
with open('/root/Desktop/classifier_result.pkl', 'wb') as handle:
    pickle.dump(raw_X, handle)

Error:

    raise TypeError, "can't pickle %s objects" % base.__name__
TypeError: can't pickle generator objects

Any help would be much appreciable.

Answer 1

raw_X = (self.token_ques(text) for text in training_data)

This is a generator. As the error says, we cannot pickle generators. Use this instead.

raw_X=[]
for text in data:
  raw_X.append(self.token_ques(text))
raw_X=tuple(raw_X)

And pickle raw_X then.

---

Edit

This works for me

import pickle

raw_X=[]
data=[1,2,3,4,5,6,2,0]
for text in data:
    raw_X.append(str(text))

print pickle.dumps(raw_X)

I'm using str() instead of your function and dumps() instead of dump().

Answer 2

Don't use a generator expression when you want to pickle data. Use a list comprehension instead, or call list() on the generator to capture all generated elements for pickling.

For example, the following works just fine:

raw_X = [self.token_ques(text) for text in training_data]
with open('/root/Desktop/classifier_result.pkl', 'wb') as handle:
    pickle.dump(raw_X, handle)

as does:

raw_X = (self.token_ques(text) for text in training_data)
with open('/root/Desktop/classifier_result.pkl', 'wb') as handle:
    pickle.dump(list(raw_X), handle)