[FEATURE] Table Acceleration Support

[ps48]

Flint Spark extension to create skipping index with OpenSearch which can accelerate query with applicable filtering condition.

Skipping Index

• User can only create, view or delete a skipping index
• One table can only have 1 skipping index for a Flint table
• Name is auto-generated for a given table and follows a pre-defined by a naming convention
• Columns can be accelerated individually using partition, value_set and min_max skipping type
• This is always auto-refreshed or manually refreshed.

Covering Index

• User can only create, view or delete a covering index
• User can have multiple covering index for a Flint table
• User can define their own name for the covering Index
• User can select columns required for creating the covering index
• This is always auto-refreshed or manually refreshed.

Materialized View

• User can only create, view or delete a materialised view
• User can have multiple materialized views for a Flint table
• User can define their own name for the materialized view
• Columns can be aggregated into metrics with different functions like count, min, max, sum and avg.
• The above metrics can be further grouped into different columns [Ex. TUMBLE() or SLIDE()]
• Final check for MVs are where clauses to restrict metrics on a given subset of data
• This is always auto-refreshed or manually refreshed.