comphead commented on code in PR #1192:
URL: https://github.com/apache/datafusion-comet/pull/1192#discussion_r1901317525
##########
native/core/src/execution/shuffle/shuffle_writer.rs:
##########
@@ -1567,17 +1585,41 @@ pub fn write_ipc_compressed<W: Write + Seek>(
let mut timer = ipc_time.timer();
let start_pos = output.stream_position()?;
- // write ipc_length placeholder
- output.write_all(&[0u8; 8])?;
+ // seek past ipc_length placeholder
+ output.seek_relative(8)?;
+
+ // write number of columns because JVM side needs to know how many
addresses to allocate
+ let field_count = batch.schema().fields().len();
+ output.write_all(&field_count.to_le_bytes())?;
let output = match codec {
CompressionCodec::None => {
+ output.write_all(b"NONE")?;
let mut arrow_writer = StreamWriter::try_new(output,
&batch.schema())?;
arrow_writer.write(batch)?;
arrow_writer.finish()?;
arrow_writer.into_inner()?
}
+ CompressionCodec::Snappy => {
+ output.write_all(b"SNAP")?;
+ let mut wtr = snap::write::FrameEncoder::new(output);
+ let mut arrow_writer = StreamWriter::try_new(&mut wtr,
&batch.schema())?;
+ arrow_writer.write(batch)?;
+ arrow_writer.finish()?;
+ wtr.into_inner()
+ .map_err(|e| DataFusionError::Execution(format!("lz4
compression error: {}", e)))?
+ }
+ CompressionCodec::Lz4Frame => {
+ output.write_all(b"LZ4_")?;
+ let mut wtr = lz4_flex::frame::FrameEncoder::new(output);
+ let mut arrow_writer = StreamWriter::try_new(&mut wtr,
&batch.schema())?;
Review Comment:
probably dumb question, looks like the writers gets created for every input.
In real life example it will be recreated for every `batch`, probably the
writer can be reused until the data stream has completed?
--
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]
