How expensive it is to load the environment to run a Python script?

I have a background python script that gets ran several thousand times a day. I'm simply running it with python foo.py. The script itself does some imports (a parsing library and sqlalchemy) and then makes a database connection, makes the parsing and saves the data to db.

I'm wondering if it adds a lot of overhead to load the python environment eac开发者_如何学Goh time the script is run?

I could make it so that the script is started once and it would have a polling loop to see if it should do something, but want to clarify that it's worth to do this.

Any input?

---

of course it adds a lot of overhead, and it would be (however negligibly) more eco-friendly to use a built-in poll or select(); but then you'd have to have a watchdog to see if it crashed, or use respawn from inittab. as long as the server load is fine, it might not be worth the effort.

---

forgot to mention, memory leaks that would be unnoticeable in a cron job can become server-eating monsters when your script runs as a daemon. you'd want to watch it carefully the first hour or two, to see if it's growing.