Expect optimise the not easy understand error message: "Can't group on '_$COUNT$_rpda_0'"

[wjl1619]

• dble version：
  5.6.29-dble-9.9.9.9-6f54f427d25b89a820752e2354eaefd01f21acde-20191025035141
• preconditions ：

drop table if exists sharding_2_t1;
create table sharding_2_t1(id int, name varchar(30) );

• configs：

schema.xml

<schema name="schema1" sqlMaxLimit="100" dataNode="dn5">
         <table name="sharding_2_t1" primaryKey="id" dataNode="dn1,dn2" rule="hash-two" />
 </schema>

• steps：
  step1. execute sql: select count(id) as clo2 from sharding_2_t1 group by clo2;

• expect result：
  1.mysql> select count(id) as clo2 from sharding_2_t1 group by clo2; ERROR 1056 (42000): Can't group on 'clo2'

• real result：
  1.mysql> select count(id) as clo2 from sharding_2_t1 group by clo2; ERROR 1003 (HY000): Can't group on '_$COUNT$_rpda_0'

• supplements：
  1.

[yanhuqing666]

delay

[yanhuqing666]

hard to repair, keep it as a limitation

[yanhuqing666]

suggest using explain to understand the error

mysql> explain select count(id) as clo2 from sharding_2_t1 group by clo2;
+-------------------+-----------------+----------------------------------------------------------------------------------------------------------------------------------------------------------+
| DATA_NODE         | TYPE            | SQL/REF                                                                                                                                                  |
+-------------------+-----------------+----------------------------------------------------------------------------------------------------------------------------------------------------------+
| dn1_0             | BASE SQL        | select COUNT(id) as `_$COUNT$_rpda_0`,COUNT(sharding_2_t1.id) as `_$COUNT$_rpda_1` from  `sharding_2_t1` GROUP BY COUNT(id) ORDER BY _$COUNT$_rpda_1 ASC |
| dn2_0             | BASE SQL        | select COUNT(id) as `_$COUNT$_rpda_0`,COUNT(sharding_2_t1.id) as `_$COUNT$_rpda_1` from  `sharding_2_t1` GROUP BY COUNT(id) ORDER BY _$COUNT$_rpda_1 ASC |
| merge_and_order_1 | MERGE_AND_ORDER | dn1_0; dn2_0                                                                                                                                             |
| aggregate_1       | AGGREGATE       | merge_and_order_1                                                                                                                                        |
| order_1           | ORDER           | aggregate_1                                                                                                                                              |
| shuffle_field_1   | SHUFFLE_FIELD   | order_1                                                                                                                                                  |
+-------------------+-----------------+----------------------------------------------------------------------------------------------------------------------------------------------------------+
6 rows in set (0.00 sec)