But I just realized that there was a observed_otus_vector.qza file in core-metrics-results folder as well.
I expected that those two files showed same value, but it didn't.
What are the differences between those two OTU tables?
The difference between these two outputs is likely due to the fact that in core-metrics you are required to provide a sampling depth which all the reads are rarefied to. Where with diversity-alpha no rarefying has been done. If you were to rarefy your table first to the same depth as the core-metrics approach, you should get the same result. Hope that helps!