duplicate column names in a table alias should be prevented.

[nicktobey]

Repro steps:

drop table if exists ab;
create table ab (a int, b int);
insert into ab values (1, 1), (2, 2);

drop table if exists bc;
create table bc (b int, c int);
insert into bc values (1, 1), (2, 2);

SELECT * FROM (SELECT * FROM ab join bc on ab.b != bc.b) as abc;

SELECT b FROM (SELECT * FROM ab join bc on ab.b != bc.b) as abc;

MySQL results:

ERROR 1060 (42S21) at line 10: Duplicate column name 'b'

go-mysql-server results:

+---+---+---+---+
| a | b | b | c |
+---+---+---+---+
| 2 | 2 | 1 | 1 |
| 1 | 1 | 2 | 2 |
+---+---+---+---+

+---+
| b |
+---+
| 2 |
| 1 |
+---+

In this example, the table alias has two columns named b, and one of them appears to be selected arbitrarily. We should be consistent with MySQL and reject this.