Currently a single python engine can possibly be reused to run multiple python data functions if necessary. Customer wants an option to always run Python data function in another newly created python engine.
The reason for this idea is that some third party modules could be much slower if being executed in the same python engine.
For example, a python data function calling below picke function takes 2 minutes only on first run but it took 20 minute (10 times slower) if run again.
p = pd.read_pickle('c:/test/df_M.pkl.zip')
Thank you Wei.
Thank you Arnaud, this request is meant for Python service only.
Hello Wei and thanks for contributing.
Is this request about data function executing through the Python service or for local execution in the Analyst client?