zhangyue19921010 commented on PR #13365:
URL: https://github.com/apache/hudi/pull/13365#issuecomment-2979971425
> @zhangyue19921010 nice work. Test with 1.7TB data, 300 columns binary copy
achived `7.7 mins` vs normal write `35 mins`; and cpu cost cut down from `90%
-> 60%`
>
> i have add some buffer for inputStream
>
> ```
> public class HoodieInputFile implements InputFile {
> @Override
> public SeekableInputStream newStream() throws IOException {
> byte[] buffer = new byte[(int) stat.getLen()];
> FSDataInputStream fsDataInputStream = fs.open(stat.getPath());
> fsDataInputStream.readFully(buffer, 0, (int) stat.getLen);
> return new DelegatingSeekableInputStream(new
ByteBufferBackedInputStream(buffer)) {
> public long getPos() ....
> public void seek....
> }
> }
> }
> ```
Thanks @xiarixiaoyao for your help in verifying the functionality and
performance of this PR. It's great to see the approximately **5x performance
improvement**. cc: @vinothchandar @danny0405 ~
--
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]
