Enhancements to be added on top of ORC Cache
- Complex predicate support in "cache table" table command
a. Cache table xxxx where c1 ='yyy' and c2 ='zzz';
- Where clause support in "drop cache" command
a. drop cache table1;
- CacheAware scheduler optimization [Quick prototyping required]
a. Primary/Secondary cache workers – Local Affinity scheduler
b. Reschedule split on non-busy (free) node but disabling caching the new node.
- Bug: ON-1257-Support all partition column types
- Show Cache
a. Amount of row data cached by each worker
b. Beautify output
- Caching Parquet, RCFile types.
- Caching non partition table.
a. Dimensions table – TPCDS dataset – [Quick prototyping required]
- Customize Cache eviction - ORC File access based eviction
- Automatic caching based on data access pattern and frequency.