Hi, David. It always depends on your use case. Did you try the following new features?
https://orc.apache.org/docs/releases.html ORC-742 LazyIO of non-filter columns ORC-577 Support row-level filtering ORC-751 Implement Predicate Pushdown in C++ Reader ORC-780 Support LZ4 Compression in C++ Writer Best, Dongjoon. On Mon, Oct 4, 2021 at 7:01 AM David <[email protected]> wrote: > Hello Gang, > > I have invested some time in squeezing out performance for ORC 1.7. > > Just curious if there are any measurable improvements out there. > > > Thanks. > > On Tue, Sep 28, 2021, 12:36 AM Dongjoon Hyun <[email protected]> > wrote: > > > Hi, All. > > > > The following is the Apache ORC 1.7.0 release and adoption status (as of > > today). > > > > 2021-09-15: Apache ORC 1.7.0 is released > > 2021-09-20: Apache Spark (dongjoon, > > https://github.com/apache/spark/pull/34045) > > 2021-09-20: Apache Iceberg (william, > > https://github.com/apache/iceberg/pull/3160) > > 2021-09-21: Apache Arrow (william, > > https://github.com/apache/arrow/pull/11194) > > 2021-09-27: Apache Druid (william, > > https://github.com/apache/druid/pull/11726) > > ON-GOING : Apache Hive (william/pgaref, > > https://github.com/apache/hive/pull/2615) > > FAILED : Apache Flink (dongjoon, > > https://github.com/apache/flink/pull/16644) > > Flink has an old fork of `PhysicalWriterImpl` based on Apache > > ORC 1.5.6. > > > > Thank you all for your efforts! > > > > Dongjoon > > >
