Improving search

ShopClass searches listing titles and descriptions through a MySQL full-text index. That is fast and needs no extra services, but it has defaults that surprise people — most often, short words that simply never match.

Words shorter than the minimum are ignored

MySQL will not index a word below a minimum length. The defaults are:

Storage engine	Variable	Default
InnoDB (normal)	innodb_ft_min_token_size	3
MyISAM (legacy)	ft_min_word_len	4

So on a default install, searching for TV, PC or BMW on a MyISAM table returns nothing at all — not “no results found”, but genuinely no match, because the word was never indexed.

If your categories are full of short model names or two-letter abbreviations, lower it.

Changing the limits

These are server settings, not application settings. Edit your MySQL or MariaDB configuration file (my.cnf, or a file under /etc/mysql/conf.d/):

[mysqld]
innodb_ft_min_token_size = 2
ft_min_word_len = 2

Restart the database for them to take effect.

Shared hosting cannot do this

Full-text tuning requires access to the database server’s configuration. If you are on shared hosting, you cannot change these — skip to when to use a real search engine.

Rebuild the index afterwards

Changing an indexing variable does not re-index existing rows. Until you rebuild, the new setting applies only to listings added after the restart — which looks exactly like the change not working.

-- InnoDB
OPTIMIZE TABLE oc_t_item_description;

-- MyISAM
REPAIR TABLE oc_t_item_description QUICK;

Replace oc_ with your actual table prefix.

Stopwords

MySQL also refuses to index very common words, from a built-in stopword list. On an English-language site this is usually what you want. On a site in another language, the English list is doing nothing useful and may be excluding real search terms — supply your own with innodb_ft_server_stopword_table, or empty the list.

Keeping search fast as the site grows

• Cache the repeated work. Search pages recompute category and location trees on every request. Object caching removes most of that.
• Watch slow queries. Enable the slow query log and look at what search actually issues before optimising anything by guesswork.
• Prune expired listings. Admin → Tools → Cleanup removes expired, spam, blocked and unactivated content. A table full of dead rows makes every search slower.

When MySQL is not enough

Full-text search in MySQL has no typo tolerance, no stemming worth relying on, no relevance tuning and no faceted scoring. Past a few hundred thousand listings, or on a site where search quality is the product, the answer is a dedicated engine — Elasticsearch, OpenSearch, Meilisearch or Typesense — driven from a plugin that indexes on the listing hooks.

That is a plugin’s job, not core’s. See the developer documentation for the hooks to index from.