vinoth chandar

@Davis-Zhang-Onehouse When you get a chance, please bring the PR description inline with the approach now implemented

@danny0405 we need a versioning schema for index storage as well, just like log blocks.. So this is useful foundation work regardless

lets review/revisit this, after 1.1 and see if this is still needed.. It may not be.

cc @vingov do you mind taking a review at this, given its a python benefiting change

In some sense, with the Spark SQL support now, python users can do custom merges? does that satisfy your requirements?

@nsivabalan are n't there any tests affected by this change? Also do we even need this post processor feature anymore? Should can the entire feature instead of just making it...

I am marking this a non-blocker.

Lets please stick to `requestedTime`, since we already have that as the state. We will rename everything consistently based on that going forward. Not the PR to be having that...