About spark todo

[turboFei]

• compatible with more datasource
• explicit cache and cache manager
• optimize spark sql operator(join, aggregate) with index
• integrate with struct streaming

索引论文: http://ilpubs.stanford.edu:8090/253/1/1997-40.pdf

[turboFei]

第四点struected streaming暂时搁置

[turboFei]

第一点除了gp，还有什么其他DataSource 需要呢？