Chungmin Lee

@rapoth [Release notes](https://github.com/microsoft/hyperspace/releases) state that whenever there are breaking changes, indexes should be "reconstructed". Does it mean that users should manually create indexes again? Because refreshAction won't work because it...

> > Does it mean that users should manually create indexes again? > > Yes > Hmm, okay. Now I see a need for a migration utility. > > Because...

I only see IndexLogEntry having such create() method which is recently added. Are there any other such classes? For IndexLogEntry, both apply() and create() are being used, so we should...

By the way, it seems you can override the apply method of the companion object since Scala 2.12.2: https://github.com/scala/scala/pull/5816. So this works: ```scala case class A(a: Int) object A {...

Implementation added as #443.

> * Will the index be bucketed? No, because the index data is expected to be very small compared to the source data and a full scan of index data...

> > > * How is data restatement handled (ie: replace /data/p1 with a new /data/p1 with different values that will require re-compute and update the min/max values in the...

I'm not sure because getIndexStat doesn't exist in the latest code base, but I think it's best to leave it to the user. If the user has 100 indexes, transposing...

> @clee704 The API name was changed during the code review :) > It's statistic API for an index. So now it's Hyperspace.index? Because it returns a DataFrame, I still...

I think this PR is good in general, but we need some clarification and write something about how and where data is stored in Hyperspace. Maybe there is one already?