Sampling of documents

I want to know about how compass samples data. It takes a sample of 1000 documents out of million documents and displays the analysis of fields that it finds in those 1000 documents.
My question- Is it possible that compass can skip/ignore some fields for analysis as it randomly takes 1000 documents and the fields which are skipped are part of those documents that compass didn’t take into its sample of 1000 ??

Yes, it’s not guaranteed that it’ll return all fields but it’ll be a good representation.

Here’s how the sampling works.