Hacker News new | past | comments | ask | show | jobs | submit login

I'm right there with you!

For disclosure I'm a minor contributor to dask so probably am a little biased.

I guess one side I probably haven't put forward though is that the memory footprint of something like dask/spark is higher because of its overheads. If you don't have scalable resources, then a polars / duckdb option would probably be your most reliable choice (I.e. the one that'll hit the fewest memory errors in the given architecture)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: