Question on batch_set_v2

[beardyFace]

I'm trying to get my head around the difference in batch_set_v2 vs batch_set, are they functionally the same but different implementations for performance reasons? My understanding is the intent is to recreate the SumTree with the inverse priorities, then sample from it - or is there more going on with how v2 sets up the tree?

Why in batch_set_v2 do you skip updating anything if the length of the indices is not even? - ( "len(ind) % 2 == 0:"). My understanding is the SumTree can have an odd number of entries.

The parameter "t" which tracks back "self.total_it" doesn't appear to be used either.