keksmd opened a new issue, #981: URL: https://github.com/apache/incubator-graphar/issues/981
### Describe the bug, including details regarding any error messages, version, and platform. `ParquetPhysicalReader` fails a `ReadRequest` whose row range covers only part of a row group when the file has no Parquet Offset Index: `UnsupportedOperationException: A partial row-group range requires a Parquet Offset Index; refusing JVM fallback.` The Arrow C++ 9.0 edge chunks in `testing/ldbc_sample` (for example `edge/person_knows_person/ordered_by_source/adj_list/part0/chunk0`) have no Offset Index, so a per-vertex adjacency range cannot be read from canonical C++-written datasets. Expected: the read returns exactly the requested rows. Where the row range cannot be applied through the Offset Index, `ReadReport` lists `ROW_RANGE` as declined rather than applied. Per the io-api contract, a declined hint changes only the report, never the rows or schema. Memory stays bounded to the row groups the range touches. ### Component(s) Java -- 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]
