Description
As for now HiveORC profile doesn't use any storage layer optimizations for LIKE operator.
Current optimization could be applied:
1) Parse first token of LIKE clause before first occurrence of "%" symbol.
2) Apply ORC filter: >=<TOKEN> AND <=<NEXT_TOKEN>
Where NEXT_TOKEN = <TOKEN>.substring(0, <TOKEN>.length() - 2) + <TOKEN>.substring(<TOKEN>.length() -2, <TOKEN>.length() - 1) + 1