Hi Smiklos,
Thanks for your interest in this issue. I am the author of this PR and now rebasing the code to the latest trunk. I have some questions: 1. Could you share how you conducted the benchmark? I want to run the full validation with all cases. 2. As you can see in the PR, it proposes three configuration alternatives. Which one do you prefer? Thanks, Dongjin +1. Sorry for the late reply. I was working on another issue. On Wed, Jan 15, 2020 at 2:01 AM smiklos <szotsmik...@gmail.com> wrote: > Hi, > > Is there any update on this? I've done performance test with Avro data > and Snappy compression. > > Setting the buffer from 32kb to 128kb brings a rough 10% decrease in > storage which is a big deal. > > I could offer working on this as well. > > Best regards, > > Miklos > > > -- *Dongjin Lee* *A hitchhiker in the mathematical world.* *github: <http://goog_969573159/>github.com/dongjinleekr <https://github.com/dongjinleekr>linkedin: kr.linkedin.com/in/dongjinleekr <https://kr.linkedin.com/in/dongjinleekr>speakerdeck: speakerdeck.com/dongjin <https://speakerdeck.com/dongjin>*
