If you estimate the new size, we certainly don't need to update any buckets to a value larger than that. This appears to be non-associative, but could be useful as a method like:
def conservative_+(kv: (K, Long)): CMS[K]
also, we could consider a variant that does this by default when merging sparse values.
See: http://dl.acm.org/citation.cfm?id=1868768
If you estimate the new size, we certainly don't need to update any buckets to a value larger than that. This appears to be non-associative, but could be useful as a method like:
also, we could consider a variant that does this by default when merging sparse values.