niyue opened a new pull request, #12812: URL: https://github.com/apache/arrow/pull/12812
This PR aims to address https://issues.apache.org/jira/projects/ARROW/issues/ARROW-16131 Currently, when writing an IPC file with multiple record batches using the `arrow::ipc::RecordBatchWriter`, the `custom_metadata` for each record batch is not saved and will be discarded silently. The current writer does provide the `AppendCustomMetadata` API internally, but it is never called. I use this API to add the custom metadata during writing and retrieve the custom metadata when reading the record batch. I am not completely sure if this is intentionally ignored or accidentally missing, but from a user's perspective, the metadata can be set via public API (see the example case in ARROW-16131) and I think it is not very friendly that this piece of data is silently discarded. Server test cases are added in the `read_write_test.cc` for verifying this change. -- 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: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org