zhangxinyao88 opened a new pull request, #3866: URL: https://github.com/apache/parquet-java/pull/3866
A leading `--conf key=value` is consumed by Hadoop's `ToolRunner` as an XML resource option, so the command can succeed without applying the property. For example, `parquet --conf parquet.avro.write-parquet-uuid=true convert ...` writes a string column instead of UUID storage. Normalize that alias to `--property` only within the leading Hadoop option prefix, then use the same entry path in `Main` and the shared CLI test helper. Preserve Hadoop's `-conf` XML resources, `-D` properties and generic option values; leave command arguments untouched. Explicit properties still override configuration-file values. Fourteen regression cases cover configuration delivery, repeated properties, precedence, Hadoop options and argument values. Two launch the actual `Main` process and inspect the converted Parquet schema. Removing the fix produces ten assertion failures, including the UUID storage regression. Validation: `./mvnw -o -B -ntp -pl parquet-cli spotless:check test` passed on Java 17: 97 tests, 95 passed, 2 skipped; Spotless and RAT passed. Full-reactor and other-JDK checks were not run. Related to #3283 and #3289. This is independent of the test-only conversion coverage in #3861. -- 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]
