huan233usc opened a new pull request, #17147: URL: https://github.com/apache/iceberg/pull/17147
Reland of #17131, which was reverted in #17141 because it conflicted with the `ParquetMetrics` `FieldStats` refactor (#17116) that landed around the same time. This version is rewritten on top of #17116's simplified metrics-test API, which passes `Metrics` directly to `assertCounts` / `assertBounds` instead of the removed `MetricsWithStats` wrapper. `TestMetrics` is the canonical per-type metrics test — it pins the value/null-count and lower/upper-bound behavior of every primitive type across the Parquet and ORC formats — but it had no geometry/geography coverage. Geo columns were given counts-only metrics in #16850 (Parquet skips the lexicographic WKB min/max, which is not meaningful for spatial data). This adds `testMetricsForGeospatialTypes` asserting that geometry and geography keep value/null counts while intentionally producing no bounds. Real spatial bounds (`geo_lower` / `geo_upper`) remain a separate follow-up. Geo values are only written by the Parquet path today, so the test is skipped for other formats (ORC). Testing: `TestParquetMetrics.testMetricsForGeospatialTypes` runs across all four format versions and passes; `TestOrcMetrics` skips it via the format assumption. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
