Please take this survey to help us learn more about how you use third party tools. Your input is greatly appreciated!
Any UDFs for cardinality estimation out there?
We use Vertica frequently for counting things. Sometimes we need an exact count, but often a little bit of error would be fine. I'm curious whether anyone has experimented with UDFs using algorithms like KMV or HyperLogLog for cardinality estimation in Vertica, and if the code is available anywhere. I'm hoping this would be useful for cutting down on memory usage when querying projections that have been optimized for a different purpose.