Hello Dear list, this is more a plpgsql dev. issue.
I wrote a plpgsql function to compute union of time range taht works : [1,4]U[3,8]U[12,14]U[16,18] ---> [1,8]U[12,14]U[16,18] It works on multiple rows. My issue is a design issue : I want to work on set of row and return set of row. I am aware I could take as input/output array of range but I don't want (memory/casting cost). Currently the function takes a cursor on a table and output a setof record. I would like that the function can blend in multiple subqueries smoothly, as WITH ( first query to get range), (query computing union ), (query using computed union ) etc. Currently I have to execute 2 sql statment : create cursor on ranges; WITH (function to compute union) , (query...) The only kind of function taking set of record as input I know of is aggregate function, but it returns only one row and the output of union can take multiple row. Any insight would be greatly appreciated. Cheers, Rémi-C
