Why? You're already provided with this information: NULL fields are not being used in the grouping set for the row. Moreover, it would seem to be an implementation- and run-time- dependent value, as there is no reason group by grouping set (a,b), (c,d) couldn't be executed in written order, or (c,d) first depending on a lot of things. Forcing the implementation to conform to a certain way of doing things is asking for someone to ask why a certain optimization isn't being performed later on. My $0.02. Jim On 03/13/2016 10:07 PM, Tom Smith wrote: > It would help if the resultset has some param to mark which is which > with the grouping sets index. > for example, the results for (a,b,c,d) would be marked as for index =0, > (b,c,d) would be index=1 > > On Sun, Mar 13, 2016 at 9:52 PM, James Keener <jim@xxxxxxxxxxxxx > <mailto:jim@xxxxxxxxxxxxx>> wrote: > > Do you want to know if a row is from the (a,b) or (c,d) group? All rows > will contain (a,b,c,d) but (a,b) will be NULL for the (c,d) grouping > sets, and vice-versa. > > Jim > > On 03/13/2016 09:45 PM, Tom Smith wrote: > > Hello: > > > > With JDBC, how can I tell which row is for which grouping sets or > rollup > > using result sets > > > > Thanks > > > > > > -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
