Re: Base Docker image caching broken in CI

Seems like it's fixed now! On Wed, 11 Jan 2023 at 15:58, Hyukjin Kwon wrote: > Hi all, > > ghcr is flaky now, so we will have to wait for a couple of days and see if > it gets fixed up soon. > See also > https://github.com/apache/spark/pull/39490#issuecomment-1378190658 > Thanks Yikun for taking

Implementation for approx_count_distinct_sketch and associated functions

Hello! I've recently wanted to write the sketches associated with the approx_count_distinct function to allow for distinct count re-aggregation. This 2019 databricks post proposes the use of spark-alchemy, and I