ianvkoeppe commented on issue #4293: Add support for an aggregation function returning serialized hyperlog… URL: https://github.com/apache/incubator-pinot/pull/4293#issuecomment-500968989 I think I can provide a more concrete example of what I meant, but there are other reasonable use cases which may be more persuasive. Use Case 1: Response Size too Large. In the case of having 10,000 IDs, let's say we group by a persons job title. There are many job titles, so for 10,000 unique IDs, we could have a very large response size which exceeds the amount of data which can be returned. We can improve performance by querying smaller groups of data by batching the IDs and then aggregating matching records client size. (In a way, this is working around not having the ability to page results when using aggregation functions). Use Case 2: Cross Column or Cross Table Aggregations. Say I have two tables; one tracks page views by member, and another tracks ad clicks by member. If I want to see unique members who visited my website and also clicked on by ad, I would not be able to query both and add their raw numbers. Assuming I'm using the same member id in both HLLs, I could merge the raw HLL responses from Pinot to achieve this.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
