> What is the performance difference we should expect (if any) between:
>
> 1. N x Model.get(key)
> i.e. making N calls to Model.get with various keys
>
> vs.
>
> 2. Model.get([key1, ... keyN])
> i.e. making a single call to Model.get with an array of N keys
>
> I'm trying to optimize out some CPU
With Method 2 the datastore can do the fetches in parallel.
The GAE team recommends method 2 for speed.
2009/1/14 boson :
>
> What is the performance difference we should expect (if any) between:
>
> 1. N x Model.get(key)
> i.e. making N calls to Model.get with various keys
>
> vs.
>
> 2. Model.g
