[20180928]如何能在11g下执行.txt

[20180928]如何能在11g下执行.txt

--//链接问的问题: http://www.itpub.net/thread-2105467-1-1.html

create table test(t_id int,t_name varchar2(50));
create table test2(t_id int,t_name varchar2(50));

insert into test values(1,'a');
insert into test values(2,'b');
insert into test values(3,'c');

insert into test2 values(1,'a');
insert into test2 values(2,'b');
insert into test2 values(3,'c');

select a.t_id, a.t_name
    from test a
         left join (  select t_id, t_name
                        from test2
                    group by t_name) b
            on a.t_id = b.t_id and a.t_name = b.t_name
   where a.t_id = 1
group by a.t_id, a.t_name;

--//同样的sql ，在10g下就能执行，在11g下就会报错 ora-00979 不是group by 表达式。红色部分看起来确实是错的，可是10g下整句
--//执行就有结果。如何能做到这个语句再11g下也能执行呢。

--//家里没有10g,使用12c测试看看,这条语句明显错误(  select t_id, t_name from test2 group by t_name),这里少写了 group by
--//t_id,t_name.

1.环境:
scott@test01p> @ ver1
port_string                    version        banner                                                                               con_id
------------------------------ -------------- -------------------------------------------------------------------------------- ----------
ibmpc/win_nt64-9.1.0           12.1.0.1.0     oracle database 12c enterprise edition release 12.1.0.1.0 - 64bit production              0

2.测试:
scott@test01p> select a.t_id,a.t_name from test a left join
  2  (select t_id,t_name from test2 group by t_name) b
  3  on a.t_id=b.t_id and a.t_name=b.t_name where a.t_id=1
  4  group by a.t_id,a.t_name;
(select t_id,t_name from test2 group by t_name) b
        *
error at line 2:
ora-00979: not a group by expression
--//12c 一样报错.

scott@test01p> show parameter feature
name                                 type                 value
------------------------------------ -------------------- ---------------
optimizer_adaptive_features          boolean              true
optimizer_features_enable            string               12.1.0.1

scott@test01p> alter session set optimizer_features_enable='10.2.0.1';
session altered.

scott@test01p> select a.t_id, a.t_name
  2      from test a
  3           left join (  select t_id, t_name
  4                          from test2
  5                      group by t_name) b
  6              on a.t_id = b.t_id and a.t_name = b.t_name
  7     where a.t_id = 1
  8  group by a.t_id, a.t_name;
      t_id t_name
---------- --------------------------------------------------
         1 a

--//ok通过.不过这句话明显错误.看看执行计划:

scott@test01p> @ dpc '' advanced
plan_table_output
-------------------------------------
sql_id  0urvg3qakxxcx, child number 0
-------------------------------------
select a.t_id, a.t_name     from test a          left join (  select
t_id, t_name                         from test2
group by t_name) b             on a.t_id = b.t_id and a.t_name =
b.t_name    where a.t_id = 1 group by a.t_id, a.t_name

plan hash value: 1211648783

-----------------------------------------------------------------------------------------------------------
| id  | operation             | name  | e-rows |e-bytes| cost (%cpu)| e-time   |  omem |  1mem | used-mem |
-----------------------------------------------------------------------------------------------------------
|   0 | select statement      |       |        |       |     9 (100)|          |       |       |          |
|   1 |  hash group by        |       |      1 |    40 |     9  (34)| 00:00:01 |  2063k|  2063k|  438k (0)|
|   2 |   view                |       |      1 |    40 |     8  (25)| 00:00:01 |       |       |          |
|   3 |    hash group by      |       |      1 |    92 |     8  (25)| 00:00:01 |  1345k|  1345k|  486k (0)|
|*  4 |     hash join outer   |       |      1 |    92 |     7  (15)| 00:00:01 |  1421k|  1421k|  722k (0)|
|*  5 |      table access full| test  |      1 |    52 |     3   (0)| 00:00:01 |       |       |          |
|*  6 |      table access full| test2 |      1 |    40 |     3   (0)| 00:00:01 |       |       |          |
-----------------------------------------------------------------------------------------------------------

query block name / object alias (identified by operation id):
-------------------------------------------------------------

   1 - sel$439afb4f
   2 - sel$3dd9cb74 / $vm_view_0@sel$439afb4f
   3 - sel$3dd9cb74
   5 - sel$3dd9cb74 / a@sel$3
   6 - sel$3dd9cb74 / test2@sel$2

outline data
-------------

  /*+
      begin_outline_data
      ignore_optim_embedded_hints
      optimizer_features_enable('10.2.0.1')
      db_version('12.1.0.1')
      all_rows
      outline_leaf(@"sel$3dd9cb74")
      merge(@"sel$2")
      outline_leaf(@"sel$439afb4f")
      outline(@"sel$f2b2f603")
      outline(@"sel$2")
      outline(@"sel$cd8351fa")
      merge(@"sel$f1d6e378")
      outline(@"sel$4")
      outline(@"sel$f1d6e378")
      merge(@"sel$1")
      outline(@"sel$3")
      outline(@"sel$1")
      no_access(@"sel$439afb4f" "$vm_view_0"@"sel$439afb4f")
      use_hash_aggregation(@"sel$439afb4f")
      full(@"sel$3dd9cb74" "a"@"sel$3")
      full(@"sel$3dd9cb74" "test2"@"sel$2")
      leading(@"sel$3dd9cb74" "a"@"sel$3" "test2"@"sel$2")
      use_hash(@"sel$3dd9cb74" "test2"@"sel$2")
      use_hash_aggregation(@"sel$3dd9cb74")
      end_outline_data
  */


predicate information (identified by operation id):
---------------------------------------------------

   4 - access("a"."t_name"="t_name" and "a"."t_id"="t_id")
   5 - filter("a"."t_id"=1)
   6 - filter("t_id"=1)

--//抽取outline.

--//编辑整理如下:

begin
dbms_sqltune.import_sql_profile(
   name => 'profile_group_error',
   description => 'sql profile created manually test',
   sql_text => q'[select a.t_id, a.t_name
    from test a
         left join (  select t_id, t_name
                        from test2
                    group by t_name) b
            on a.t_id = b.t_id and a.t_name = b.t_name
   where a.t_id = 1
group by a.t_id, a.t_name]',
   profile => sqlprof_attr(
'ignore_optim_embedded_hints',
q'[optimizer_features_enable('10.2.0.1')]',
q'[db_version('12.1.0.1')]',
'all_rows',
'outline_leaf(@"sel$3dd9cb74")',
'merge(@"sel$2")',
'outline_leaf(@"sel$439afb4f")',
'outline(@"sel$f2b2f603")',
'outline(@"sel$2")',
'outline(@"sel$cd8351fa")',
'merge(@"sel$f1d6e378")',
'outline(@"sel$4")',
'outline(@"sel$f1d6e378")',
'merge(@"sel$1")',
'outline(@"sel$3")',
'outline(@"sel$1")',
'no_access(@"sel$439afb4f" "$vm_view_0"@"sel$439afb4f")',
'use_hash_aggregation(@"sel$439afb4f")',
'full(@"sel$3dd9cb74" "a"@"sel$3")',
'full(@"sel$3dd9cb74" "test2"@"sel$2")',
'leading(@"sel$3dd9cb74" "a"@"sel$3" "test2"@"sel$2")',
'use_hash(@"sel$3dd9cb74" "test2"@"sel$2")',
'use_hash_aggregation(@"sel$3dd9cb74")'
),
   replace => true,
   force_match => true
);
end;
/

scott@test01p> show parameter feature
name                                 type                 value
------------------------------------ -------------------- --------------
optimizer_adaptive_features          boolean              true
optimizer_features_enable            string               12.1.0.1

scott@test01p> select a.t_id, a.t_name
  2      from test a
  3           left join (  select t_id, t_name
  4                          from test2
  5                      group by t_name) b
  6              on a.t_id = b.t_id and a.t_name = b.t_name
  7     where a.t_id = 1
  8  group by a.t_id, a.t_name;
      t_id t_name
---------- --------------------------------------------------
         1 a

--//换1个数值,select变成select看看.

scott@test01p> select a.t_id, a.t_name
  2      from test a
  3           left join (  select t_id, t_name
  4                          from test2
  5                      group by t_name) b
  6              on a.t_id = b.t_id and a.t_name = b.t_name
  7     where a.t_id = 2
  8  group by a.t_id, a.t_name;
      t_id t_name
---------- --------------------------------------------------
         2 b

--//ok解决,最好还是叫开发改代码..