How Can I Update Global Variables in the Parent Process When Using Multiprocessing in Python?

Multiprocessing's Inability to Update Global Variables in Parent Process

In multiprocessing, the parent and child processes exist in separate Python instances, with their own memory and global variables. Changes made to shared global variables in child processes are not accessible to the parent process. Additionally, the constraints imposed by serialization protocols such as pickle prohibit the transfer of unpickleable objects.

Addressing the Challenge

One potential solution is to find a way to store data using ctype objects, allowing references to the memory to be passed to child processes. However, this approach is not straightforward.

Specific to the LORR Object

The non-pickleability of the LORR object poses a further challenge. To overcome this, one could consider subclassing LORR and implementing __setstate__ and __getstate__ methods for pickling. Alternatively, it may be possible to extract the necessary data from the LORR instance and pass it as a string via Queues, eliminating the need for object serialization. This approach involves sending message instructions to child processes and receiving responses containing the required data.

The above is the detailed content of How Can I Update Global Variables in the Parent Process When Using Multiprocessing in Python?. For more information, please follow other related articles on the PHP Chinese website!