數據庫實踐丨MySQL多表join分析

時間 2020-08-06

標籤數據庫實踐 mysql join 分析欄目 SQL 简体版

原文原文鏈接

摘要：在數據庫查詢中,每每會須要查詢多個表的數據,好比查詢會員信息同時查詢關於這個會員的訂單信息,若是分語句查詢的話,效率會很低,就須要用到join關鍵字來連表查詢了。

Join並行

Join並行1. 多表join介紹2. 多表Join的方式不使用Join buffer使用Join buffer3. Join執行流程（老執行器）mysql

1. 多表join介紹

JOIN子句用於根據兩個或多個表之間的相關列來組合它們。例如：sql

Orders：數據庫

Customers：segmentfault

SELECT Orders.OrderID, Customers.CustomerName, Orders.OrderDate
FROM Orders
INNER JOIN Customers ON Orders.CustomerID=Customers.CustomerID;

2. 多表Join的方式

Hash join使用新執行器實現，在這裏不作討論api

MySQL支持的都是Nested-Loop Join，以及它的變種。oop

不使用Join buffer

a) Simple Nested-Loop

對r表的每一行，完整掃描s表，根據r[i]-s[i]組成的行去判斷是否知足條件，並返回知足條件的結果給客戶端。this

mysql> show create table t1;
+-------+----------------------------------------------------------------------------------------------------------------+
| Table | Create Table                                                                                                   |
+-------+----------------------------------------------------------------------------------------------------------------+
| t1    | CREATE TABLE `t1` (
 `id` int(11) NOT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+-------+----------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> show create table t3;
+-------+--------------------------------------------------------------------------------------------------------------------+
| Table | Create Table                                                                                                       |
+-------+--------------------------------------------------------------------------------------------------------------------+
| t3    | CREATE TABLE `t3` (
 `id` int(11) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+-------+--------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> explain select /*+ NO_BNL() */ * from t1, t3 where t1.id = t3.id;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | t1    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    2 |   100.00 | NULL        |
|  1 | SIMPLE      | t3    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    2 |    50.00 | Using where |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-------------+
2 rows in set, 1 warning (0.00 sec)

b) Index Nested-Loop

對r表的每一行，先根據鏈接條件去查詢s表索引，而後回表查到匹配的數據，並返回知足條件的結果給客戶端。lua

mysql> show create table t2;
+-------+---------------------------------------------------------------------------------------------------------------------------------------+
| Table | Create Table                                                                                                                          |
+-------+---------------------------------------------------------------------------------------------------------------------------------------+
| t2    | CREATE TABLE `t2` (
 `id` int(11) NOT NULL,
 KEY `index1` (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+-------+---------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> explain select * from t1, t2 where t1.id = t2.id;
+----+-------------+-------+------------+------+---------------+--------+---------+------------+------+----------+-------------+
| id | select_type | table | partitions | type | possible_keys | key    | key_len | ref        | rows | filtered | Extra       |
+----+-------------+-------+------------+------+---------------+--------+---------+------------+------+----------+-------------+
|  1 | SIMPLE      | t1    | NULL       | ALL  | NULL          | NULL   | NULL    | NULL       |    2 |   100.00 | NULL        |
|  1 | SIMPLE      | t2    | NULL       | ref  | index1        | index1 | 4       | test.t1.id |    1 |   100.00 | Using index |
+----+-------------+-------+------------+------+---------------+--------+---------+------------+------+----------+-------------+
2 rows in set, 1 warning (0.00 sec)

使用Join buffer

a) Block Nested Loop

從r表讀取一部分數據到join cache中，當r表數據讀完或者join cache滿後，作join操做。spa

JOIN_CACHE_BNL::join_matching_records(){
 do {
   //讀取s表的每一行
   qep_tab->table()->file->position(qep_tab->table()->record[0]);
   //針對s的每一行，遍歷join buffer
   for(each record in join buffer) {
     get_record();
     rc = generate_full_extensions(get_curr_rec());
     //若是不符合條件，直接返回
     if (rc != NESTED_LOOP_OK) return rc;
   }
 } while(!(error = iterator->Read()))
}

mysql> explain select  * from t1, t3 where t1.id = t3.id;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------------------------------------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra                                              |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------------------------------------------+
|  1 | SIMPLE      | t1    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    2 |   100.00 | NULL                                               |
|  1 | SIMPLE      | t3    | NULL       | ALL  | NULL          | NULL | NULL    | NULL |    2 |    50.00 | Using where; Using join buffer (Block Nested Loop) |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+----------------------------------------------------+
2 rows in set, 1 warning (0.00 sec)

b) Batched Key Access

從r表讀取一部分數據到join cache中，s表中記錄r表被鏈接的列的值做爲索引，查詢全部符合條件的索引，而後將這些符合條件的索引排序，而後統一回表查詢記錄。code

其中，對於每個cached record，都會有一個key，經過這個key去s表掃描所需的數據。

dsmrr_fill_buffer(){
 while((rowids_buf_cur < rowids_buf_end) &&
       !(res = h2->handler::multi_range_read_next(&range_info))){
   //下壓的index條件
   if (h2->mrr_funcs.skip_index_tuple &&
       h2->mrr_funcs.skip_index_tuple(h2->mrr_iter, curr_range->ptr))
     continue;
   memcpy(rowids_buf_cur, h2->ref, h2->ref_length);
 }
 varlen_sort(
     rowids_buf, rowids_buf_cur, elem_size,
     [this](const uchar *a, const uchar *b) { return h->cmp_ref(a, b) < 0; });
}

dsmrr_next(){
 do{
   if (rowids_buf_cur == rowids_buf_last) {
     dsmrr_fill_buffer();
   }
   // first match
   if (h2->mrr_funcs.skip_record &&
       h2->mrr_funcs.skip_record(h2->mrr_iter, (char *)cur_range_info, rowid))
     continue;
   res = h->ha_rnd_pos(table->record[0], rowid);
   break;
 } while(true);
}

JOIN_CACHE_BKA::join_matching_records(){
 while (!(error = file->ha_multi_range_read_next((char **)&rec_ptr))) {
   get_record_by_pos(rec_ptr);
   rc = generate_full_extensions(rec_ptr);
     if (rc != NESTED_LOOP_OK) return rc;
 }
}

mysql> show create table t1;
+-------+-------------------------------------------------------------------------------------------------------------------------------------------------+
| Table | Create Table                                                                                                                                    |
+-------+-------------------------------------------------------------------------------------------------------------------------------------------------+
| t1    | CREATE TABLE `t1` (
 `f1` int(11) DEFAULT NULL,
 `f2` int(11) DEFAULT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+-------+-------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> show create table t2;
+-------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table | Create Table                                                                                                                                                                                   |
+-------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| t2    | CREATE TABLE `t2` (
 `f1` int(11) NOT NULL,
 `f2` int(11) NOT NULL,
 `f3` char(200) DEFAULT NULL,
 KEY `f1` (`f1`,`f2`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+-------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> explain SELECT /*+ BKA() */ t2.f1, t2.f2, t2.f3 FROM t1,t2 WHERE t1.f1=t2.f1 AND t2.f2 BETWEEN t1.f1 and t1.f2 and t2.f2 + 1 >= t1.f1 + 1;
+----+-------------+-------+------------+------+---------------+------+---------+-------------+------+----------+---------------------------------------------------------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref         | rows | filtered | Extra                                                         |
+----+-------------+-------+------------+------+---------------+------+---------+-------------+------+----------+---------------------------------------------------------------+
|  1 | SIMPLE      | t1    | NULL       | ALL  | NULL          | NULL | NULL    | NULL        |    3 |   100.00 | Using where                                                   |
|  1 | SIMPLE      | t2    | NULL       | ref  | f1            | f1   | 4       | test1.t1.f1 |    7 |    11.11 | Using index condition; Using join buffer (Batched Key Access) |
+----+-------------+-------+------------+------+---------------+------+---------+-------------+------+----------+---------------------------------------------------------------+
2 rows in set, 1 warning (0.00 sec)

c) Batched Key Access(unique)

與Batched Key Access不一樣的是，r中的列是s的惟一索引，在r記錄寫入join cache的時候，會記錄一個key的hash table，僅針對不一樣的key去s表中查詢。（疑問，爲何只有unique的時候才能用這種方式？不是unique的話，s表中可能會掃描出多條數據，也能夠用這種方式去處理，減小s表的重複掃描）。
JOIN_CACHE_BKA_UNIQUE::join_matching_records(){
 while (!(error = file->ha_multi_range_read_next((char **)&key_chain_ptr))) {
   do(each record in chain){
     get_record_by_pos(rec_ptr);
     rc = generate_full_extensions(rec_ptr);
       if (rc != NESTED_LOOP_OK) return rc;
     }
 }
}

mysql> show create table city;
+-------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table | Create Table                                                                                                                                                                                                                                                                                                                                       |
+-------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| city  | CREATE TABLE `city` (
 `ID` int(11) NOT NULL AUTO_INCREMENT,
 `Name` char(35) NOT NULL DEFAULT '',
 `Country` char(3) NOT NULL DEFAULT '',
 `Population` int(11) NOT NULL DEFAULT '0',
 PRIMARY KEY (`ID`),
 KEY `Population` (`Population`),
 KEY `Country` (`Country`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+-------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

mysql> show create table country;
+---------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Table   | Create Table                                                                                                                                                                                                                                                                                                                                                      |
+---------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| country | CREATE TABLE `country` (
 `Code` char(3) NOT NULL DEFAULT '',
 `Name` char(52) NOT NULL DEFAULT '',
 `SurfaceArea` float(10,2) NOT NULL DEFAULT '0.00',
 `Population` int(11) NOT NULL DEFAULT '0',
 `Capital` int(11) DEFAULT NULL,
 PRIMARY KEY (`Code`),
 UNIQUE KEY `Name` (`Name`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci |
+---------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.01 sec)

mysql> EXPLAIN SELECT city.Name, country.Name FROM city,country WHERE city.country=country.Code AND  country.Name LIKE 'L%' AND city.Population > 100000;
+----+-------------+---------+------------+-------+--------------------+---------+---------+--------------------+------+----------+--------------------------------------------------------------+
| id | select_type | table   | partitions | type  | possible_keys      | key     | key_len | ref                | rows | filtered | Extra                                                        |
+----+-------------+---------+------------+-------+--------------------+---------+---------+--------------------+------+----------+--------------------------------------------------------------+
|  1 | SIMPLE      | country | NULL       | index | PRIMARY,Name       | Name    | 208     | NULL               |    1 |   100.00 | Using where; Using index                                     |
|  1 | SIMPLE      | city    | NULL       | ref   | Population,Country | Country | 12      | test1.country.Code |    1 |   100.00 | Using where; Using join buffer (Batched Key Access (unique)) |
+----+-------------+---------+------------+-------+--------------------+---------+---------+--------------------+------+----------+--------------------------------------------------------------+
2 rows in set, 1 warning (0.01 sec)

3. Join執行流程（老執行器）

sub_select <--------------------------------------------+
 | -> iterator::read() // 讀一行數據                    |
 | -> evaluate_join_record()  //檢查這行數據是否符合條件 |
 | -> next_select() ---+                               |
                       |                               |
sub_select_op  <--------+                               |
 | -> op->put_record() // 前表數據寫入join cache        |
   | -> put_record_in_cache()                          |
   | -> join->record()                                 |
     | -> join_matching_records()                      |
       | -> (qep_tab->next_select)(join, qep_tab + 1, 0) // 繼續調用next_select
   | -> end_send()

點擊關注，第一時間瞭解華爲雲新鮮技術~