Breaking changes
- Now
max_n,max_n_by,min_n, andmin_n_byreturn an empty result instead of raising an error when used on an empty stream of data. - 1D aggregations now throw an error when any number would loose precision (53 bits) instead of returning wrong data.
What's Changed
- Use ORG_AUTOMATION_TOKEN instead of API_TOKEN_GITHUB_PACKAGE by @svenklemm in #902
- Update readme by @atovpeko in #903
- Mark time_weight (and its rollup) parallel unsafe by @syvb in #882
- Adjust ci setup for recent ubuntu versions by @svenklemm in #908
- Update pgrx to v0.18.0 by @surister in #910
- Set superuser and trusted in extension control file by @svenklemm in #916
- Update CI/Build versions of Ubuntu and Rockylinux by @surister in #914
- Apply Rustc and Clippy fixes by @Dvdandrades in #918
- Fix
max_nandmax_n_bywhen used on empty tables by @Dvdandrades in #915 - Fix
min_nandmin_n_bywhen used on empty tables by @Dvdandrades in #917 - Release/prepare 1.23.0 by @surister in #920
- Fix inverted is_sorted check in ASAP timevector smoothing by @Dvdandrades in #921
- Add guard for large BIGINT values in 1D aggregations. by @surister in #905
New Contributors
- @atovpeko made their first contribution in #903
- @Dvdandrades made their first contribution in #918
Full Changelog: 1.22.0...1.23.0