liukun4515 commented on code in PR #1855: URL: https://github.com/apache/arrow-rs/pull/1855#discussion_r903740918
########## arrow/src/ipc/writer.rs: ########## @@ -922,20 +1021,67 @@ fn write_array_data( } /// Write a buffer to a vector of bytes, and add its ipc::Buffer to a vector +/// From <https://github.com/apache/arrow/blob/6a936c4ff5007045e86f65f1a6b6c3c955ad5103/format/Message.fbs#L58> +/// Each constituent buffer is first compressed with the indicated +/// compressor, and then written with the uncompressed length in the first 8 +/// bytes as a 64-bit little-endian signed integer followed by the compressed +/// buffer bytes (and then padding as required by the protocol). The +/// uncompressed length may be set to -1 to indicate that the data that +/// follows is not compressed, which can be useful for cases where +/// compression does not yield appreciable savings. fn write_buffer( buffer: &Buffer, buffers: &mut Vec<ipc::Buffer>, arrow_data: &mut Vec<u8>, offset: i64, + compression_codec: &CompressionCodecType, ) -> i64 { - let len = buffer.len(); - let pad_len = pad_to_8(len as u32); - let total_len: i64 = (len + pad_len) as i64; - // assert_eq!(len % 8, 0, "Buffer width not a multiple of 8 bytes"); - buffers.push(ipc::Buffer::new(offset, total_len)); - arrow_data.extend_from_slice(buffer.as_slice()); - arrow_data.extend_from_slice(&vec![0u8; pad_len][..]); - offset + total_len + let origin_buffer_len = buffer.len(); + let mut _compression_buffer = Vec::<u8>::new(); + let (data, uncompression_buffer_len) = match compression_codec { + CompressionCodecType::NoCompression => { + // this buffer_len will not used in the following logic + // If we don't use the compression, just write the data in the array + (buffer.as_slice(), origin_buffer_len as i64) + } + CompressionCodecType::Lz4Frame | CompressionCodecType::Zstd => { + if (origin_buffer_len as i64) == LENGTH_EMPTY_COMPRESSED_DATA { + (buffer.as_slice(), 0) + } else { + #[cfg(any(feature = "zstd,lz4", test))] + compression_codec + .compress(buffer.as_slice(), &mut _compression_buffer) + .unwrap(); Review Comment: @viirya comments like above reply Maybe I can resolve this option or compile issue in the next issue or pull request. In this pull request, we can focus on the protocal of IPC compression -- 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]
