相同的數據庫，不同的mysql：查詢時間執行的巨大差異

Question

我有兩個MySQL安裝，一個在Xampp上，另一個在Mac上。機器在硬件上類似。
數據庫是相同的，從一個DUMP文件導入到每個MySQL中。

我發佈此兩個EXPLAINS EXTENDED：

EXPLAIN EXTENDED (ON MAC) 

+----+--------------------+--------------+------------+--------+--------------------------+----------+---------+---------------------------------------+------+----------+---------------------------------------------------------------+ 
| id | select_type  | table  | partitions | type | possible_keys   | key  | key_len | ref         | rows | filtered | Extra               | 
+----+--------------------+--------------+------------+--------+--------------------------+----------+---------+---------------------------------------+------+----------+---------------------------------------------------------------+ 
| 1 | PRIMARY   | little_tasks | NULL  | ref | post_id,meta_key   | meta_key | 767  | const         | 620 |  0.01 | Using where; Using temporary; Using filesort; Start temporary | 
| 1 | PRIMARY   | exp_baby  | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  0.39 | Using where             | 
| 1 | PRIMARY   | certs  | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  0.39 | Using where             | 
| 1 | PRIMARY   | radius  | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  3.00 | Using where             | 
| 1 | PRIMARY   | exp_years | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  1.00 | Using where             | 
| 1 | PRIMARY   | remun_max | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  1.00 | Using where             | 
| 1 | PRIMARY   | edu   | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  1.00 | Using where             | 
| 1 | PRIMARY   | start_date | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  3.04 | Using where             | 
| 1 | PRIMARY   | moment  | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  3.04 | Using where             | 
| 1 | PRIMARY   | lat   | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  3.53 | Using where             | 
| 1 | PRIMARY   | lon   | NULL  | ref | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.little_tasks.post_id | 12 |  3.53 | Using where             | 
| 1 | PRIMARY   | p   | NULL  | eq_ref | PRIMARY,type_status_date | PRIMARY | 8  | helpeeit_helpee2.little_tasks.post_id | 1 | 24.26 | Using where; End temporary         | 
| 3 | DEPENDENT SUBQUERY | NULL   | NULL  | NULL | NULL      | NULL  | NULL | NULL         | NULL |  NULL | Impossible WHERE            | 
| 2 | SUBQUERY   | spoken_lang | NULL  | ref | post_id,meta_key   | meta_key | 767  | const         | 620 |  0.02 | Using where             | 
+----+--------------------+--------------+------------+--------+--------------------------+----------+---------+---------------------------------------+------+----------+---------------------------------------------------------------+ 
14 rows in set, 2 warnings (4 min 18,58 sec) 

EXPLAIN EXTENDED (ON WINDOWS/XAMPP) 

+----+--------------------+--------------+----------------+--------------------------+----------+---------+-------------------------------------+------+----------+----------------------------------------------+ 
| id | select_type  | table  | type   | possible_keys   | key  | key_len | ref         | rows | filtered | Extra          | 
+----+--------------------+--------------+----------------+--------------------------+----------+---------+-------------------------------------+------+----------+----------------------------------------------+ 
| 1 | PRIMARY   | radius  | ref   | post_id,meta_key   | meta_key | 767  | const        | 619 | 100.00 | Using where; Using temporary; Using filesort | 
| 1 | PRIMARY   | p   | eq_ref   | PRIMARY,type_status_date | PRIMARY | 8  | helpeeit_helpee2.radius.post_id  | 1 | 100.00 | Using where         | 
| 1 | PRIMARY   | exp_years | ref   | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.p.ID    | 5 | 100.00 | Using where         | 
| 1 | PRIMARY   | start_date | ref   | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.exp_years.post_id | 5 | 100.00 | Using where         | 
| 1 | PRIMARY   | moment  | ref   | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.start_date.post_id | 5 | 100.00 | Using where         | 
| 1 | PRIMARY   | lat   | ref   | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.radius.post_id  | 5 | 100.00 | Using where         | 
| 1 | PRIMARY   | lon   | ref   | post_id,meta_key   | post_id | 8  | helpeeit_helpee2.lat.post_id  | 5 | 100.00 | Using where         | 
| 8 | DEPENDENT SUBQUERY | edu   | index_subquery | post_id,meta_key   | post_id | 8  | func        | 5 | 100.00 | Using where         | 
| 7 | DEPENDENT SUBQUERY | certs  | index_subquery | post_id,meta_key   | post_id | 8  | func        | 5 | 100.00 | Using where         | 
| 6 | DEPENDENT SUBQUERY | little_tasks | index_subquery | post_id,meta_key   | post_id | 8  | func        | 5 | 100.00 | Using where         | 
| 5 | DEPENDENT SUBQUERY | exp_baby  | index_subquery | post_id,meta_key   | post_id | 8  | func        | 5 | 100.00 | Using where         | 
| 4 | DEPENDENT SUBQUERY | remun_max | index_subquery | post_id,meta_key   | post_id | 8  | func        | 5 | 100.00 | Using where         | 
| 3 | DEPENDENT SUBQUERY | NULL   | NULL   | NULL      | NULL  | NULL | NULL        | NULL |  NULL | Impossible WHERE        | 
| 2 | DEPENDENT SUBQUERY | spoken_lang | index_subquery | post_id,meta_key   | post_id | 8  | func        | 5 | 100.00 | Using where         | 
+----+--------------------+--------------+----------------+--------------------------+----------+---------+-------------------------------------+------+----------+----------------------------------------------+ 
14 rows in set, 1 warning (0.03 sec) 

你可以自己4分鐘看到的執行時間上的MAC而在Windows/XAMPP

只有 0.03秒

這裏是查詢條件：

SELECT DISTINCT p.ID,p.post_author, lat.meta_value AS lat, lon.meta_value AS lon, radius.meta_value AS radius, start_date.meta_value AS date_interval, moment.meta_value AS day_moment, exp_years.meta_value AS exp_years 
FROM `hlp_posts` p 

INNER JOIN hlp_postmeta AS lat ON (p.ID = lat.post_id AND lat.meta_key = "hlp_latitude") 
INNER JOIN hlp_postmeta AS lon ON (p.ID = lon.post_id AND lon.meta_key = "hlp_longitude") 
INNER JOIN hlp_postmeta AS radius ON (p.ID = radius.post_id AND radius.meta_key = "hlp_resume_range_distance") 
INNER JOIN hlp_postmeta AS start_date ON (p.ID = start_date.post_id AND start_date.meta_key = "hlp_resume_date") 
INNER JOIN hlp_postmeta AS moment ON (p.ID = moment.post_id AND moment.meta_key = "hlp_resume_availability") 
INNER JOIN hlp_postmeta AS exp_years ON (p.ID = exp_years.post_id AND exp_years.meta_key = 'resume_years_of_exp' AND (exp_years.meta_value >= 4)) 

WHERE 
p.post_type = 'babysitter' 
AND p.post_status = 'publish' 

AND (
p.ID IN (SELECT spoken_lang.post_id FROM hlp_postmeta AS spoken_lang WHERE spoken_lang.meta_key = 'wpjobus_resume_languages' AND (spoken_lang.meta_value LIKE '%Inglese%' AND spoken_lang.meta_value LIKE '%Francese%' AND spoken_lang.meta_value LIKE '%Spagnolo%' AND spoken_lang.meta_value LIKE '%Tedesco%')) 
OR 
p.ID IN (SELECT native_lang.post_id FROM hlp_postmeta AS native_lang WHERE native_lang.meta_key = 'wpjobus_resume_native_language' AND native_lang.meta_key = 'wpjobus_resume_native_language' AND (native_lang.meta_value = 'Inglese' AND native_lang.meta_value = 'Francese' AND native_lang.meta_value = 'Spagnolo' AND native_lang.meta_value = 'Tedesco')) 
) 

AND p.ID IN (SELECT remun_max.post_id FROM hlp_postmeta AS remun_max WHERE remun_max.meta_key = 'wpjobus_resume_remuneration' AND remun_max.meta_value <= 5) 
AND p.ID IN (SELECT exp_baby.post_id FROM hlp_postmeta AS exp_baby WHERE exp_baby.meta_key = 'wpjobus_resume_skills' AND (exp_baby.meta_value LIKE '%i:0%' AND exp_baby.meta_value LIKE '%i:1%' AND exp_baby.meta_value LIKE '%i:2%' AND exp_baby.meta_value LIKE '%i:3%')) 
AND p.ID IN (SELECT little_tasks.post_id FROM hlp_postmeta AS little_tasks WHERE little_tasks.meta_key = 'hlp_resume_little_tasks' AND (little_tasks.meta_value LIKE '%i:0%' AND little_tasks.meta_value LIKE '%i:1%' AND little_tasks.meta_value LIKE '%i:2%' AND little_tasks.meta_value LIKE '%i:3%' AND little_tasks.meta_value LIKE '%i:4%' AND little_tasks.meta_value LIKE '%i:5%' AND little_tasks.meta_value LIKE '%i:6%')) 
AND p.ID IN (SELECT certs.post_id FROM hlp_postmeta AS certs WHERE certs.meta_key = 'hlp_resume_prof_edu' AND (certs.meta_value LIKE '%i:0%' AND certs.meta_value LIKE '%i:2%' AND certs.meta_value LIKE '%i:3%' AND certs.meta_value LIKE '%i:7%' AND certs.meta_value LIKE '%i:11%')) 
AND p.ID IN (SELECT edu.post_id FROM hlp_postmeta AS edu WHERE edu.meta_key = 'hlp_resume_edu' AND (edu.meta_value >= 3)) 

ORDER BY `p`.`ID` DESC 

我真的無法解釋這是怎麼可能的。運行實際查詢之前

SET optimizer_search_depth = 0; 

：

Answer 1

我發現張貼的問題作爲一個bug here的解決方案。
也從5.6更新到5.7可以提高性能。