How to save all python objects from working directory to a file

Question

I need to find the way to save all objects (or at least dataframes) to one place, outside of working directory. I assume Python keeps all objects in memory not on disk, so I'm looking for the way of exporting all objects from current session. It can be pickle, does not matter until you can read it to a different Python session

Answer 1

If you want to save a dataframe in another directory, try :

df.to_csv('foldername/filename.csv')

Answer 2

If I understood your need, you want to backup your session. If that is the case, here is a solution using pickle. Kr.

First solution is:

import pickle

def is_picklable(obj):
    try:
        pickle.dumps(obj)
    except Exception:
        return False
    return True

bk = {}
for k in dir():
    obj =  globals()[k]
    if is_picklable(obj):
        try:
            bk.update({k: obj})
        except TypeError:
            pass

# to save session
with open('./your_bk.pkl', 'wb') as f:
    pickle.dump(bk, f)

    
# to load your session

with open('./your_bk.pkl', 'rb') as f:
    bk_restore = pickle.load(f)

***Second solution is with dill. You might have error if in your workspace, there are some unpicklable objects ***:

import dill

dill.dump_session('./your_bk_dill.pkl')
#to restore session:
dill.load_session('./your_bk_dill.pkl')

Third option go with shelve package:

import shelve

bk = shelve.open('./your_bk_shelve.pkl','n')
for k in dir():
    try:
        bk[k] = globals()[k]
    except Exception:
        pass
bk.close()

# to restore
bk_restore = shelve.open('./your_bk_shelve.pkl')
for k in bk_restore:
    globals()[k]=bk_restore[k]
    tmp[k] = bk_restore[k]
bk_restore.close()

Check and let's know about your trial.

Credits: The second and third solution are nearly a shameless copy/paste from those two links belows. I adapted the handling of errors as the original answer will lead to error for pickling of module.
dill solution
shelve solution