Optimize take_bits for non-null indices

[Dandandan]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
Take bits uses a normal iterator on the indices. This makes it slower, as it has to check each value on being non-null or not.

Describe the solution you'd like
We can write a fast path for non-nulls.

Describe alternatives you've considered
Not doing this.

Additional context