I'm writing my own /Group by/ operator (non-hashed implementation) and I'm
currently looking on a good data structure to store the result groups. There
are two requirement in my mind: the table need to be efficiently expanded
and the tuples within each group need to be sorted.
I carefully examined the implementation of standard /Group by/ (both hashed
and sorted implementation), the closest I got is to leverage parts of the
/TupleHashTable/ but before going through this plan, I need to make sure I'm
not overlooking an existing data structure.
Thanks
--
View this message in context: http://postgresql.1045698.n5.nabble.com/TupleTable-like-data-structure-tp5748432.html
Sent from the PostgreSQL - hackers mailing list archive at Nabble.com.