oracle運算式之“ora-0079”不是group by運算式分析（未完待續。。。）__oracle

with tab as (select t.object_pid,         t.mesh,         round(x, 8) x,         round(y, 8) y,         round(z, 3) z,         id    from had_object_box_pole t, table(sdo_util.GetVertices(t.geometry)) g),tab1 as (select t.object_pid,         t.mesh,         t.x,         t.y,         t.z,         t.id,         count(*) over(partition by t.object_pid, t.x, t.y, t.z) num,         row_number() over(partition by t.object_pid, t.x, t.y, t.z order by t.id) sq    from tab t   where t.id > 1),tab2 as (select t.object_pid,         t.mesh,         t.x,         t.y,         t.z,         t.id,         t.num,         t.sq,         (t.x || ',' || t.y || ',' || t.z) str    from tab1 t   where num > 1),tab3 as (select object_pid,         t.id,         mesh,        substr(sys_connect_by_path(str,'|'),2) strVal,         wm_concat(t.id) strid    from tab2 t   where connect_by_isleaf = 1   start with sq = 1  connect by sq = PRIOR sq + 1         and object_pid = PRIOR object_pid         and sq < =(select max(sq) from tab2) group by object_pid,t.id,mesh)select * from tab3 t3

在進行資料庫操作時，進行tab3的選取時出現如題的錯誤，起初按照網上說的在select 清單項目中出現的列必須出現在group by後面（彙總函式除外），但是不能正確，找到的替代件決辦法還是採用substr(sys_connect_by_path)將id進行拼串。但是這個問題解決不了是心中的一塊結啊。