I have a table like this:
CREATE TABLE `products` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`name` varchar(512) NOT NULL,
`description` text,
PRIMARY KEY (`id`),
) ENGINE=InnoDB AUTO_INCREMENT=38 DEFAULT CHARSET=utf8;
and one like this:
CREATE TABLE `product_variants` (
`id` int(11) unsigned NOT NULL AUTO_INCREMENT,
`product_id` int(11) unsigned NOT NULL,
`product_code` varchar(255) NOT NULL,
PRIMARY KEY (`id`),
UNIQUE KEY `product_code` (`product_code`),
KEY `product_variant_product_fk` (`product_id`),
CONSTRAINT `product_variant_product_fk` FOREIGN KEY (`product_id`) REFERENCES `products` (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=1037 DEFAULT CHARSET=utf8;
and an SQL statement like this
SELECT p.id AS id, p.name AS name, p.description AS description, pv.id AS product_variant_id, pv.product_code AS product_code
FROM products p
INNER JOIN product_variants pv ON pv.product_id = p.id
ORDER BY p.name ASC
LIMIT 300 OFFSET 0;
which if I explain gives me this:
+----+-------------+-------+------+----------------------------+----------------------------+---------+---------+--------+----------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+----------------------------+----------------------------+---------+---------+--------+----------------+
| 1 | SIMPLE | p | ALL | PRIMARY | NULL | NULL | NULL | 993658 | Using filesort |
| 1 | SIMPLE | pv | ref | product_variant_product_fk | product_variant_product_fk | 4 | db.p.id | 1 | |
+----+-------------+-------+------+----------------------------+----------------------------+---------+---------+--------+----------------+
2 rows in set (0.00 sec)
For a million rows, this is pretty slow. I've tried adding an index on products.name with:
ALTER TABLE products ADD INDEX `product_name_idx` (name(512));
which gives this:
mysql> show indexes from products;
+----------+------------+------------------+--------------+-----------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | Index_comment |
+----------+------------+------------------+--------------+-----------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
| products | 0 | PRIMARY | 1 | id | A | 993658 | NULL | NULL | | BTREE | | |
| products | 1 | product_manf_fk | 1 | manufacturer_id | A | 18 | NULL | NULL | YES | BTREE | | |
| products | 1 | product_name_idx | 1 | name | A | 201 | 255 | NULL | | BTREE | | |
+----------+------------+------------------+--------------+-----------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
3 rows in set (0.00 sec)
I think that the Sub_part column shows the prefix that has been in indexed (in bytes), as described on this page.
When I re-explain the query, I get:
+----+-------------+-------+------+----------------------------+----------------------------+---------+---------+--------+----------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+------+----------------------------+----------------------------+---------+---------+--------+----------------+
| 1 | SIMPLE | p | ALL | PRIMARY | NULL | NULL | NULL | 993658 | Using filesort |
| 1 | SIMPLE | pv | ref | product_variant_product_fk | product_variant_product_fk | 4 | db.p.id | 1 | |
+----+-------------+-------+------+----------------------------+----------------------------+---------+---------+--------+----------------+
2 rows in set (0.00 sec)
which looks like the new index is not being used. As described on this page, indexes will not be used for sorting if they are prefix indexes. In fact if I truncate the data with:
alter table products modify `name` varchar(255) not null;
The explain gives:
+----+-------------+-------+-------+----------------------------+----------------------------+---------+----------------------------------------------+------+-------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+-------+----------------------------+----------------------------+---------+----------------------------------------------+------+-------+
| 1 | SIMPLE | p | index | PRIMARY | product_name_idx | 767 | NULL | 300 | |
| 1 | SIMPLE | pv | ref | product_variant_product_fk | product_variant_product_fk | 4 | oh_2c98c233_69fe_4f06_ad0d_fe6f85a5beac.p.id | 1 | |
+----+-------------+-------+-------+----------------------------+----------------------------+---------+----------------------------------------------+------+-------+
which I think backs that up. However, it says on this page that InnoDB tables can have up to 767 bytes of index. If the length is in bytes, why does it refuse to have more than 255? If it's in characters, how is it deciding the length of each UTF-8 character? Is it just assuming 3?
Also, am using this version of MySQL:
mysql> select version();
+------------+
| version() |
+------------+
| 5.5.27-log |
+------------+
1 row in set (0.00 sec)
The index key prefix length limit is 767 bytes for InnoDB tables that use the REDUNDANT or COMPACT row format. For example, you might hit this limit with a column prefix index of more than 191 characters on a TEXT or VARCHAR column, assuming a utf8mb4 character set and the maximum of 4 bytes for each character.dev.mysql.com/doc/refman/8.0/en/innodb-limits.html – Disdainful