Hi All, I'm interested in attending the next sync to discuss the possibility of not deprecating summary files, for the use case described in this previous thread. While Delta.io is in the state that it is without an open source implementation of partition pruning, summary files are a pretty easy fix to this performance problem if users are able to maintain them properly.
In addition to those who discussed on the list, I also had a member of the team at microsoft message me about this approach, so I think the interest from several community members could justify keeping this code around. https://mail-archives.apache.org/mod_mbox/parquet-dev/202010.mbox/%3CCAFGcCdUSDTRVfFNUJXh_tL-RPjxkQCZpEnr_s-sow8rPMg56rg%40mail.gmail.com%3E Jason Altekruse
