I think if possible landing "Make `ColumnMetaData.path_in_schema` optional" would be good to avoid a second release soon after.
On Thu, May 28, 2026 at 9:00 AM Gang Wu <[email protected]> wrote: > Hi all, > > I would like to start a discussion about a parquet-format 2.13.0 release. > > It has been about 9 months since 2.12.0 was released on August 28, 2025. > > Since then, we have had some meaningful updates, especially: > > - [PR #514](https://github.com/apache/parquet-format/pull/514) > PARQUET-2249: Introduce IEEE 754 total order & NaN-counts > - [PR #565](https://github.com/apache/parquet-format/pull/565) GH-506: > Clarify field ordering for partially shredded object union > - [PR #560](https://github.com/apache/parquet-format/pull/560) Update > wording in Geospatial spec > - [PR #562](https://github.com/apache/parquet-format/pull/562) GH-561: > variant schema examples to use `(VARIANT(1))` > > There are also some open PRs that may be worth waiting for, especially: > > - [PR #564](https://github.com/apache/parquet-format/pull/564) GH-563: > Make `ColumnMetaData.path_in_schema` optional > > It looks promising to me, and there is already related implementation > work in arrow-rs, parquet-java, and arrow-cpp. (Thanks Ed!) > > What do people think? > > Thanks, > Gang >
