Create an Aggregator that returns the estimate size, not the HLL approximate data structure itself.
Create an Aggregator that returns the estimate size, not the HLL approximate data structure itself. This is convenient, but cannot be combined later with another unique count like an HLL could.
is the log of the size the HLL.
Give an approximate set size (not the HLL) based on inputs of Array[Byte] see HyperLogLog.bitsForError for a size table based on the error see SetSizeHashAggregator for a version that uses exact sets up to a given size
This creates an HLL for type K, that uses (2^bits) bytes to store
Give a HyperLogLog Aggregator that have the given error.
Give a HyperLogLog Aggregator that have the given error. It is up to you, using bitsForError, to see if the size is still practical for your application.
0.016 (1.6%), 4 KB 0.006 (0.6%), 32 KB 0.002 (0.2%), 256 KB 0.001 (0.1%), 1024 KB
Cutting the error in half takes 4x the size.
This object makes it easier to create Aggregator instances that use HLL