在多表联接的场景中，优化器的一个很重要的任务是决定各个表之间的联接顺序（Join Order），因为不同的联接顺序会影响中间结果集的大小，进而影响到计划整体的执行代价。

    为了减少执行计划的搜索空间和计划执行的内存占用，OceanBase 数据库优化器在生成联接顺序时主要考虑左深树的联接形式。下图展示了左深树、右深树和多支树的计划形状。

    数

    OceanBase 数据库联接顺序的生成采用了 System-R 的动态规划算法，考虑的因素包括每一个表可能的访问路径、Interesting Order、联接算法（NESTED-LOOP、BLOCK-BASED NESTED-LOOP 或者 SORT-MERGE 等）以及不同表之间的联接选择率等。

    给定 N 个表的联接，OceanBase 数据库生成联接顺序的方法如下:

    1. 为每一个基表生成访问路径，保留代价最小的访问路径以及有所有有 Interesting Order 的路径。一个路径 如果具有 Interesting Order，它的序能够被后续的算子使用。

    2. 生成所有表集合的大小为 i (1 < i <= N) 的计划。 OceanBase 数据库一般只考虑左深树，表集合大小为 i 的计划可以由一个表集合大小为 i 的计划和一个基表的计划组成。OceanBase 数据库按照这种策略，考虑了所有的联接算法以及 Interesting Order 的继承等因素把所有表集合大小为 i 的计划生成。这里也只是保留代价最小的计划以及所有具有 Interesting Order 的计划。

    同时，OceanBase 数据库提供了 HINT 机制 /*+LEADING(table_name_list)*/去控制多表联接的顺序。

    如下例所示，开始选择的联接顺序是先做 t1、t2 的 JOIN 联接，然后再和 t3 做 JOIN 联接；如果用户希望先做 t2、t3 的 JOIN 联接，然后再和 t1做 JOIN 联接，则可以使用 HINT /*+LEADING(t2,t3,t1)*/去控制；如果用户希望先做 t1、t3 的 JOIN 联接，然后再和 t2 做 JOIN 联接，则可以使用 HINT /*+LEADING(t1,t3,t2)*/去控制。

    1. obclient>CREATE TABLE t1(c1 INT, c2 INT, PRIMARY KEY(c1));
    2. Query OK, 0 rows affected (0.31 sec)
    4. obclient>CREATE TABLE t2(c1 INT, c2 INT, PRIMARY KEY(c1));
    5. Query OK, 0 rows affected (0.33 sec)
    7. obclient>CREATE TABLE t3(c1 INT, c2 INT, PRIMARY KEY(c1));
    8. Query OK, 0 rows affected (0.44 sec)
    10. obclient>EXPLAIN SELECT * FROM t1,t2,t3 WHERE t1.c1 = t2.c2 AND t2.c1 = t3.c2;
    11. +-----------------------------------------------------------------+
    12. | Query Plan                                                                              |
    13. +-----------------------------------------------------------------+
    14. | =======================================
    15. |ID|OPERATOR    |NAME|EST. ROWS|COST  |
    16. ---------------------------------------
    17. |0 |HASH JOIN   |    |98010    |926122|
    18. |1 | TABLE SCAN |T3  |100000   |61860 |
    19. |2 | HASH JOIN  |    |99000    |494503|
    20. |3 |  TABLE SCAN|T1  |100000   |61860 |
    21. |4 |  TABLE SCAN|T2  |100000   |61860 |
    22. =======================================
    24. Outputs & filters: 
    25. -------------------------------------
    26.   0 - output([T1.C1], [T1.C2], [T2.C1], [T2.C2], [T3.C1], [T3.C2]), filter(nil), 
    27.       equal_conds([T2.C1 = T3.C2]), other_conds(nil)
    28.   1 - output([T3.C2], [T3.C1]), filter(nil), 
    29.       access([T3.C2], [T3.C1]), partitions(p0)
    30.   2 - output([T1.C1], [T1.C2], [T2.C1], [T2.C2]), filter(nil), 
    31.       equal_conds([T1.C1 = T2.C2]), other_conds(nil)
    32.   3 - output([T1.C1], [T1.C2]), filter(nil), 
    33.       access([T1.C1], [T1.C2]), partitions(p0)
    34.   4 - output([T2.C2], [T2.C1]), filter(nil), 
    35.       access([T2.C2], [T2.C1]), partitions(p0)
    37. obclient>EXPLAIN SELECT /*+LEADING(t2,t3,t1)*/* FROM t1,t2,t3 WHERE t1.c1 = t2.c2
    38.         AND t2.c1 = t3.c2;
    39. +-----------------------------------------------------------------+
    40. | Query Plan                                                                              |
    41. +-----------------------------------------------------------------+
    42. | ========================================
    43. |ID|OPERATOR    |NAME|EST. ROWS|COST   |
    44. ----------------------------------------
    45. |0 |HASH JOIN   |    |98010    |1096613|
    46. |1 | HASH JOIN  |    |99000    |494503 |
    47. |2 |  TABLE SCAN|T2  |100000   |61860  |
    48. |3 |  TABLE SCAN|T3  |100000   |61860  |
    49. |4 | TABLE SCAN |T1  |100000   |61860  |
    50. ========================================
    52. Outputs & filters: 
    53. -------------------------------------
    54.   0 - output([T1.C1], [T1.C2], [T2.C1], [T2.C2], [T3.C1], [T3.C2]), filter(nil), 
    55.       equal_conds([T1.C1 = T2.C2]), other_conds(nil)
    56.   1 - output([T2.C1], [T2.C2], [T3.C1], [T3.C2]), filter(nil), 
    57.       equal_conds([T2.C1 = T3.C2]), other_conds(nil)
    58.   2 - output([T2.C2], [T2.C1]), filter(nil), 
    59.       access([T2.C2], [T2.C1]), partitions(p0)
    60.   3 - output([T3.C2], [T3.C1]), filter(nil), 
    61.       access([T3.C2], [T3.C1]), partitions(p0)
    62.   4 - output([T1.C1], [T1.C2]), filter(nil), 
    63.       access([T1.C1], [T1.C2]), partitions(p0)
    65. obclient>EXPLAIN SELECT /*+LEADING(t1,t3,t2)*/* FROM t1,t2,t3 WHERE t1.c1 = t2.c2 
    66.        AND t2.c1 = t3.c2;
    67. +-----------------------------------------------------------------+
    68. | Query Plan                                                                              |
    69. +-----------------------------------------------------------------+
    70. | =============================================================
    71. |ID|OPERATOR                   |NAME|EST. ROWS  |COST       |
    72. -------------------------------------------------------------
    73. |0 |HASH JOIN                  |    |98010      |53098071243|
    74. |1 | NESTED-LOOP JOIN CARTESIAN|    |10000000000|7964490204 |
    75. |2 |  TABLE SCAN               |T1  |100000     |61860      |
    76. |3 |  MATERIAL                 |    |100000     |236426     |
    77. |4 |   TABLE SCAN              |T3  |100000     |61860      |
    78. |5 | TABLE SCAN                |T2  |100000     |61860      |
    79. =============================================================
    81. Outputs & filters: 
    82. -------------------------------------
    83.   0 - output([T1.C1], [T1.C2], [T2.C1], [T2.C2], [T3.C1], [T3.C2]), filter(nil), 
    84.       equal_conds([T1.C1 = T2.C2], [T2.C1 = T3.C2]), other_conds(nil)
    85.   1 - output([T1.C1], [T1.C2], [T3.C1], [T3.C2]), filter(nil), 
    86.       conds(nil), nl_params_(nil)
    87.   2 - output([T1.C1], [T1.C2]), filter(nil), 
    88.       access([T1.C1], [T1.C2]), partitions(p0)
    89.   3 - output([T3.C1], [T3.C2]), filter(nil)
    90.   4 - output([T3.C2], [T3.C1]), filter(nil), 
    91.       access([T3.C2], [T3.C1]), partitions(p0)
    92.   5 - output([T2.C2], [T2.C1]), filter(nil), 
    93.       access([T2.C2], [T2.C1]), partitions(p0)