Are you using protobuf python cpp extension or pure python? We do have a design of "Optimized Python NumPy Bindings for Repeated Fields" for cpp extension , but the implementation breaks too many users so it was blocked. I will check with the owner to go if we can go forward 在2022年6月15日星期三 UTC-7 14:59:14<Daniel> 写道：
> I am having the following problem: > > I have implemented a protobuff to send data from a golang client to a > python server using a grpc stream. The data I am sending needs to be loaded > quickly on the python server and processed. This data is a composite field > with the largest part being a repeated sint array defined as such: *repeated > sint32 array = 4 [packed=true];* > > This field contains around 18,000,000 entries and when I try to load these > into my code using the following line: *data = np.array(array_obj, > dtype=np.int8) *this process takes around 1.5 SECONDS. I have tried > alternative methods of first reading the data as a list, which is also not > faster, using copy=False in numpy... I just want to access the memory where > these values are stored... > > I would like to try something such as numba o Cython, but both of those > would require me to implement the complete container type stored here > https://github.com/protocolbuffers/protobuf/blob/main/python/google/protobuf/internal/containers.py > > . Is there some way this process could be accelerated? > > Thankful for any help > -- You received this message because you are subscribed to the Google Groups "Protocol Buffers" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/protobuf/2efd86bb-e0f6-4140-bed7-5dfe65e1e3aan%40googlegroups.com.
