mbutrovich commented on issue #1798: URL: https://github.com/apache/datafusion-comet/issues/1798#issuecomment-2920626829
> Here is a related PR to Spark that helped fix the problem in Spark RAPIDS. > > [apache/spark@52e3cf9](https://github.com/apache/spark/commit/52e3cf9ff50b4209e29cb06df09b1ef3a18bc83b) This looks similar to what we just discussed: tweak the canonicalized plan to encode which operator originated the shuffle files. Downside: might do the shuffle twice. Upside: shouldn't get incompatible exchanges anymore. Long term we could consider the optimizer rule that I mentioned of "if there are equivalent shuffles in a plan, but one could be native and the other not, consider follow back for the reuse benefit rather than run two different stages to generate the same shuffle data." -- 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...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org