Ryan Blue

@omervk, you might want to have a look at #52, which adds single-table transactions. That would enable you to revert a commit and replace its data in the same commit....

I'd say change the current test classes. The Avro module has its own TestReadProjection because it was easier to copy it than to mess with the build to have tests...

There's already a location in the metadata that we use to build paths. We could just use that.

Removed distinct counts in 75088f6875fc8d3cc4c3af38899742de1b919abf.

Can you give an example of properties you're trying to set here? We can cache file systems in HadoopTableOperations, but most of the systems I've worked on use this pattern...

Okay, how about adding the support you're talking about to HadoopTableOperations and opening a PR? That would unblock you because you'd have the caching level you need and we could...

Also, why do all of the properties include "spark"?

Properties set through Spark wouldn't need to be specific to Spark. You might use the same ones as session properties in Presto.

Agreed, we should plan on passing the table properties through so all of the write configuration happens there.

I think what you're trying to accomplish would be done a little differently. I understand the term "partitioning function" to mean the partition transformations that are part of a partition...