Spark managed spill in columnar sort

[zhouyuan]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
Sort should support Spill under lower memory. This could reduce memory footprint and allow to run on bigger dataset

Describe the solution you'd like
As Spark did, Sort should respect the memory size and spill to disk under low memory

Describe alternatives you've considered
N/A

Additional context
N/A