Spark read path

[jayhan94]

• Support a simple sequential scan (full table read) without optimizations.
• Support projection.
• Support predicate push down.
• Support partition prune.
• Support time travel.
• and so on.