Should MEDS datasets write _metadata or _common_metadata files?

[mmcdermott]

I don't use these, and am certainly not suggesting these would be high priority, but according to this:

https://arrow.apache.org/docs/python/parquet.html#writing-metadata-and-common-metadata-files

some frameworks use them to better parse sharded/partitioned datasets.

[EthanSteinberg]

I think we should consider adding this if it becomes an issue for someone, but to not bother until then. These datasets are small enough that I'm skeptical there would be that much benefit (those files are mainly helpful for thousands of shards IIRC)

[mmcdermott]

sounds good to me. I'll leave it up to you to close this issue or leave it open; either way hopefully folks who want this functionality (if any such exist) will be able to find it and comment on it in the future to indicate interest.