Can someone provide an example of how to use parallel table function in oracle pl/sql. We need to run massive queries for 15 years and combine the result.

SELECT * 
  FROM Table(TableFunction(cursor(SELECT * FROM year_table))) 

...is what we want effectively. The innermost select will give all the years, and the table function will take each year and run massive query and returns a collection. The problem we have is that all years are being fed to one table function itself, we would rather prefer the table function being called in parallel for each of the year. We tried all sort of partitioning by hash and range and it didn't help.

Also, can we drop the keyword PIPELINED from the function declaration? because we are not performing any transformation and just need the aggregate of the resultset.