Is there a better way via multiprocessing?

Question

1. I am using python multiprocessing to process files. Last processed file record is stored in a dict A i.e. dict_A = {'file1_xx': '8-04-22', 'file2_xx': '8-04-22', 'file3_xx': '8-04-22', 'file4_xx': '8-04-22'}
2. Files directory is scanned, filenames with last modified date are stored in dict_test. Files recorded in both dicts are compared for new files: i.e. compare each file last modified date i.e file1_xx against the last processed date in dict_A. There's a condition which will update the dict_A if the file last modified date is greater than last processed date per single file.
3. I am facing issues as the dictionary is not updated after the files are processed.
4. Ideally the dict_A should be updated with the latest modified date per file of same category. This dict_A is then uploaded to PostgreSQL db through sqlalchemy.

def compare_rec(i):
    a = dict_A[i]
    b = dict_test[i]
    if a >= b:
        print("none")
    else:
        lock.acquire()
        print("found")
        a = b
        lock.release()

def init(l):
    global lock
    lock = l

if __name__ == '__main__':
    file_cat=['a', 'b', 'c', 'd']
    dict_A={'a': '10', 'b': '10', 'c': '10', 'd': '10'}
    dict_test={'a': '11', 'b': '11', 'c': '11', 'd': '11'}
    l = multiprocessing.Lock()
    pool = multiprocessing.Pool(initializer=init, initargs=(l,))
    pool.map(compare_rec, file_cat)
    pool.close()
    pool.join()

Answer 1

Processes don't share variables.

In function I would use return to send filename and date back to main process

if ...:
    return i, a  
else:
    return i, b  

main thread should get results from all processes

results = pool.map(compare_rec, file_cat)

and it should update dictonary

dict_A.update(results)

---

Full code:

import multiprocessing

def compare_rec(key):
    print('key:', key)
    
    a = dict_A[key]
    b = dict_test[key]
    
    if a >= b:
        print("none", key, a)
        return key, a   
    else:
        print("found:", key, b)
        return key, b
    
if __name__ == '__main__':

    file_cat  = ['a', 'b', 'c', 'd']
    dict_A    = {'a': '10', 'b': '10', 'c': '10', 'd': '10'}
    dict_test = {'a': '11', 'b': '11', 'c': '11', 'd': '11'}

    pool = multiprocessing.Pool()
    
    results = pool.map(compare_rec, file_cat)
    print(results)
    
    print('before:', dict_A)
    dict_A.update(results)
    print('after :', dict_A)
    
    pool.close()
    pool.join()