Utilize secondary indexes with "arbitrary" expressions

Secondary indexes currently cannot be utilized beyond simple range and point lookups. There's a missed opportunity to exploit secondary indexes for arbitrary expressions. 

Example:

SELECT e.name, e.salary

FROM Employee e

WHERE lowercase(e.name) LIKE "%smith"

 If we have an index on Employee.name (call it name_idx), it won't be utilized.

However, we can do a scan on the secondary index and apply the expression on secondary key(s):

... <The rest is similar to what's currently done for unnest-map> ...

select like(lowercase($$name), "%smith")

  data-scan []<-[$$name, $$PK] <- Employee.name_idx