aakshintala opened a new issue, #221: URL: https://github.com/apache/arrow-java/issues/221
### Describe the bug, including details regarding any error messages, version, and platform. The `Buffer Alignment and Padding` [section](https://arrow.apache.org/docs/format/Columnar.html#buffer-alignment-and-padding) of the Arrow Columnar Format Specification allows padding to be either 8-bytes or 64-bytes with 64-bytes being preferred for SIMD performance reasons. I have a custom implementation of the Arrow Columnar Format (serialization and deserialization only as we operate on an entirely different columnar format) where I pad all of my data buffers to 64-bytes when serializing. This works fine with the C++, Python and Rust libraries (and I think with the Go library as well) as they all use the `offset` value from the flatbuffer `Buffer` message when calculating where to read the next body buffer from... The Java library however, seems to assume that the padding will always be 8-bytes and just pads the size of the `n-1`th buffer to 8-bytes when calculating the offset for the `n`th buffer. This assumption obviously breaks when the padding size is anything other than 8-bytes... Would it be acceptable to bring the Java library in-line with the remaining implementations, i.e., have it respect the offset from the flatbuffer metadata? ### 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: issues-unsubscr...@arrow.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
