Question

I have a simple QUERY with an INNER JOIN. The tables that are joined have 1500 en 2500 rows. Both tables have a primary key. The join is not done on this key.

The result set of this query has 1500rows.
When I run this query it takes 20 seconds… (AN ETERNITY)

SELECT so.order_id
FROM   shop_orders so
       INNER JOIN shop_order_payments sop
          ON so.order_code = sop.order_code
WHERE  sop.status = 9
ORDER  BY date_created DESC

When I then add the ” LIMIT 1500 ” at the end , the query takes only 1sec.
(When I remove it it takes again 20 seconds so it’s not the cache…)

1. Why is this ?
2. And is there a fast way to find out how many rows this query has as a result?

note: I run this query in a php script with the mysql() function )

*update1: The result of the EXPLAIN

id   select_type     table   type    possible_keys   key     key_len     ref     rows    Extra
1   SIMPLE  sop ALL NULL    NULL    NULL    NULL    1153    Using where; Using temporary; Using filesort
1   SIMPLE  so  ALL NULL    NULL    NULL    NULL    2529    Using where

• Update 2:

I ran DESC as requested in the comments

payment_id - int(11)    NO  PRI     auto_increment
timestamp - timestamp   NO      CURRENT_TIMESTAMP   
order_code - varchar(100)NO         
acceptance - varchar(50) NO         
amount  -  varchar(20)  NO          

So only a primary index shows up, this is not enough as an index?

Answer 1

The EXPLAIN tells that the query is using no index. Thus it performs fulltable scan causing it to be slow.

try to run this statements,

ALTER TABLE shop_orders ADD INDEX td_idx (order_code);
ALTER TABLE shop_order_payments ADD INDEX td2_idx (order_code);

and run the SELECT query again.

UPDATE 1

Your table has an index but on column payment_id which in this case isn’t used since the join is using column order_code. You must also define an index on that column.