hi, i am looking into json support of the upcoming 9.2 release and have a question about the row_to_json function. is there a way to specify the column aliases of a nested composite row without creating a custom type ? here is an example: select row_to_json(t1) from ( select 'joe' as username, row(1, 'prj1') project ) t1; the output is: {"username":"joe","project":{"f1":1,"f2":"prj1"}} i want the output to be: {"username":"joe","project":{"project_id":1,"project_name":"prj1"}} i thought of something like this: select row_to_json(t1) from ( select 'joe' as username, (select * from (values(1, 'prj1')) as project(project_id, project_name)) as project ) t1; but this fails with "ERROR: subquery must return only one column". any help would be really appreciated. regards jan zimmek - Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general