Description

This PR refactors the code in preparation for new extensions for IncrementalIndex (i.e., #10001).
It encapsulates common IncrementalIndex functionality in IncrementalIndex.java as much as possible.
In addition, it avoids exposing the dimensions storage outside of IncrementalIndexRow. That is, fetch each dimension directly without fetching the entire dimension array.

The modifications to IncrementalIndex includes:

1. Changed the getMetric methods to accept IncrementalIndexRow instead of int rowOffset, because OakIncrementalIndex does not keep a mapping from row index to an actual row.
   This does not affect the other implementations' performance because in all the cases these methods are used, the caller already had an IncrementalIndexRow object.
2. Move CachingColumnSelectorFactory to IncrementalIndex

The modifications to IncrementalIndexRow allow lazy evaluation of off-heap keys, without adding an overhead to the on-heap keys case. We add/modify the following methods:

1. Changed public Object[] getDims() to public Object getDim(int index)
2. Add public int getDimsLength()
3. Add public boolean isDimNull(int index)
4. Add public IndexedInts getIndexedDim(int index, @Nullable IndexedInts cachedIndexedInts)

getIndexedDim() purpose is to support a lazy-evaluation version of an indexed ints dimension instead of the array of integers that is returned by getDim().
The modified implementation in StringDimensionIndexer, utilize this method and query using IndexedInts instead of the array. This allows our implementation (OakIncrementalIndex) to use the lazy-evaluation approach, while the other implementations use the existing integer array, without performance degradation.

Key changed/added classes in this commit

• Changed:
  • IncrementalIndex
  • IncrementalIndexRow
  • OnheapIncrementalIndex
  • StringDimensionIndexer

This PR has:

• been self-reviewed.
• added documentation for new or modified features or behaviors.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.