Subquery is running faster than join?

问题

I am trying to identify value from one column which does exist in another column in same table

Sub query

SELECT DISTINCT `Wear it With - Outfits 1` 
  FROM `product list` 
 WHERE `Wear it With - Outfits 1` NOT IN (SELECT `sku` 
                                            FROM `product list`)

...returns result in 2.7287sec

I tried to replace sub query by left join

   SELECT DISTINCT table1.`Wear it With - Outfits 1`
     FROM `product list` as table1 
LEFT JOIN `product list` as table2 ON table1.`Wear it With - Outfits 1`=table2.sku
     WHERE table2.sku IS NULL 
       AND table1.`Wear it With - Outfits 1` IS NOT NULL

...which returns result in 5.7651 sec

Normally joins return results much faster. So i believe i did something funny in my query? But can not find any reason why my subquery is running faster than

回答1:

The statement 'Normally joins return results much faster.' is silly, especially without reference to any particular database system.

Many factors go into determining the performance of a particular query. You can use the EXPLAIN tool in whatever database product you're using to determine exactly why the sub-query is preferable in this case (hint: it likely has to do with the use of the keyword DISTINCT).

回答2:

The main reason is that your left join statement is not optimized . The WHERE condition WHERE table2.sku IS NULL and table1.Wear it With - Outfits 1 IS NOT NULL can waste a lot of time especially in this case. You should optimize table2 before you left join. PS:The number of records in table2 should reach a considerable amount.

来源：https://stackoverflow.com/questions/7830725/subquery-is-running-faster-than-join

标签

mysql

sql

join

subquery