MySQL優化之in、exists、join

一、inner join 、 in 、exists

    explain
    select a.id
    from application as a
    where exists(
      select 1
      from dispatch_app_history as d
      where d.bomber_id = 165 and d.application_id = a.id
    );

 

   

    explain
    select a.id
    from application as a
    where a.id in (
      select d.application_id
      from dispatch_app_history as d
      where d.bomber_id = 165
    );

 

    

    explain
    select a.id
    from application as a
    inner join dispatch_app_history as d on d.application_id = a.id
    where d.bomber_id = 165;

分析：子查詢需要application_id來關聯外部表application，因為需要application_id字段，所以MySQL認為無法先執行這個子查詢，而對application表進行全表查詢。

結論：子查詢、join查詢具體性能怎么樣需要根據實際情況決定

 

 

二、not in、not exists

    1、執行速度  

     explain
     select *
     from dispatch_app as d
     where d.application_id not in(
        select dh.application_id
        from dispatch_app_history as dh
     );                       

     

 

    

     explain
     select *
     from dispatch_app as d
     where not exists (
        select 1
        from dispatch_app_history as dh
        where d.application_id = dh.application_id
     )

   

  結論：不考慮其他情況，通常情況下not exists的執行效率要高於not in

 

    2、條件中含有null

     select *
     from bomber as b
     where not exists(
        select 1
        from dispatch_app_history as d
        where d.partner_id = b.partner_id
     )
     group by b.partner_id;

 

     

     select *
     from bomber as b
     where b.partner_id not in(
        select distinct d.partner_id
        from dispatch_app_history as d
     );

  

結論：對於not in，條件中有null值的時候會直接停止執行返回null結果。對於not exists，條件中有null時會清除null后執行。in、exists的執行與not existst相同