davidlghellin commented on PR #19908: URL: https://github.com/apache/datafusion/pull/19908#issuecomment-4282638831
> > Might need someone from comet or sail to chip in, they might be more familiar with how concepts map between DataFusion and Spark > > In Sail we currently have two PRs exploring this topic. I believe the key part is the ~~next_gaussian~~ implementation. @Jefffrey @cht42 quick correction to my earlier comment — `nextGaussian` is for `randn()` / normal distributions, not `uniform()`. For `uniform` the relevant primitives are `nextDouble()` / `nextInt(bound)` from `XORShiftRandom`, which are already bit-compatible in `sail-function/.../xorshift.rs`. The `uniform` PR ([#1074](https://github.com/lakehq/sail/pull/1074)) wires those in directly and produces values bit-exact vs Spark JVM for any given seed. -- 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]
