|
As I said, since downloads happen in short lived worker processes, there is absolutely no way that the speed of downloading can be affected by the size of the parent job. You will get overall faster results by splitting into multiple jobs because of the parallel downloads but that has the effect of hammering servers and so is not something I recommend doing.
And given that bulk downloading metadata for large numbers of books is not something that needs to happen frequently, I suggest you just queue up your job overnight and stop worrying about it.
|