Range bytes counts and keys count

[amirouche]

It will be useful to have:

1. Count of bytes:
   a. The count of bytes in keys in a given range
   b. The count of bytes in key-value pairs in a given range
2. The count of keys in given range

That allows to control how much memory will be used (e.g. slice transactions into manageable ranges), but also avoid the need to implement statistics at user level, to be able to optimize query plans.

[amirouche]

Having estimates instead of exact value could be enough.