Optimize hashing, null-free fast path in percentile_cont, median #23953

Is your feature request related to a problem or challenge?

The non-distinct PercentileContAccumulator and MedianAccumulator used the default SipHash hasher for their internal HashMaps. This is slow; switching to foldhash would be significantly faster.

Also, we can add a null-free fast path to update_batch and retract_batch in both accumulators.

Both ideas copied from recent work on percentile_cont(DISTINCT) in #23946

Describe the solution you'd like

No response

Describe alternatives you've considered

No response

Additional context

No response