我有一张这样的表：
// table_A
+--------+---------------+------------+
| id | short-text | long-text |
+--------+---------------+------------+
int(11) varchar(200) text // <= type
注1： sort-text或long-text中的一个始终为NULL。从来没有两个都没有包含数据。像这样：
+--------+---------------+------------+
| id | short-text | long-text |
+--------+---------------+------------+
| 1 |some character | NULL |
| 2 |some character | NULL |
| 3 |some character | NULL |
| 4 |some character | NULL |
| 5 |some character | NULL |
| 6 |NULL | some text |
| 7 |some character | NULL |
| 8 |some character | NULL |
| 9 |some character | NULL |
| 10 |some character | NULL |
| 11 |some character | NULL |
| 12 |NULL | some text |
+--------+---------------+------------+
注2：正如您在上表示例中看到的那样，总共有5行long-text为NULL，然后是1行，short-text为{ {1}}。
现在我的问题是，上面的结构更好还是只创建一列（NULL）而不是这两列？像这样的东西：
type = text这是包含数据的表：
// table_B
+--------+---------------------+
| id | short-and-long-text |
+--------+---------------------+
int(11) text // <= type
嗯，哪一个更有效率和优化？ +--------+---------------------+
| id | short-and-long-text |
+--------+---------------------+
| 1 | some character |
| 2 | some character |
| 3 | some character |
| 4 | some character |
| 5 | some character |
| 6 | some text |
| 7 | some character |
| 8 | some character |
| 9 | some character |
| 10 | some character |
| 11 | some character |
| 12 | some text |
+--------+---------------------+
或table_A？
注3：我总是按ID查找。