Search #

This page is for developers who build the search of a site. Exponential ships a search engine that keeps its index in the site's database and needs nothing else; Solr through the ezfind extension is the option for sites that need relevance tuning, facets or large volumes. The content model decides what is searchable in both.

Contents #

1. Two engines
2. What is indexed
3. When content is indexed
4. The search page
5. Searching from templates
6. Searching from PHP
7. Rebuilding the index
8. Choosing for a project

Two engines #

settings/solr.ini [SolrSettings] Enabled=false is shipped; ./console exp:solr status says whether a Solr is configured and answers (Search backends). This repository does not state which Solr versions eZ Find needs; take that from the extension's own repository.

What is indexed #

Every attribute whose class attribute is searchable (is_searchable) is indexed through its datatype; untick it for fields that should never match (internal notes, codes). The object name, class and section are indexed too, so results can be limited by them. Objects are found only by users who may read them: the search checks content/read like a fetch (Sections, roles and policies).

When content is indexed #

At publishing, unless [SearchSettings] DelayedIndexing is enabled (or classbased with DelayedIndexingClassList[]). Then publishing only queues the object, and the indexcontent.php cronjob of the default part indexes it later. Publishing large objects is faster that way, but a site with delayed indexing depends on cron: without it new content never appears in search (Delayed indexing and cron, Cronjobs).

The search page #

/content/search is the search view; /content/advancedsearch the advanced form. The result template is content/search.tpl of your design chain: override it to match the site, keep the field name SearchText, and page with the shipped navigator (Search results template). Limits of the built-in engine are settings in [SearchSettings]:

Searching from templates #

{def $result = fetch( 'content', 'search', hash(
        'text', $search_text,
        'subtree_array', array( $node.node_id ),
        'limit', 10,
        'offset', $view_parameters.offset ) )}
{foreach $result.SearchResult as $hit}
    {node_view_gui content_node=$hit view='line'}
{/foreach}
<p>{$result.SearchCount} results</p>
{undef $result}

The parameters of the search fetch (kernel/content/function_definition.php): text, subtree_array, offset, limit, publish_timestamp, publish_date, section_id, class_id, class_attribute_id, ignore_visibility, limitation and sort_by; class_id takes numeric class IDs. Use it for a search box limited to one part of the site; for lists without a search text, a list or tree fetch is cheaper (Fetching content in templates).

Searching from PHP #

eZSearch::search( $text, $params ) runs the configured engine, so the same code works with the built-in engine and with eZ Find. Check the parameters your engine accepts in its class before relying on them (kernel/classes/ezsearch.php, the engine class named by SearchEngine).

Rebuilding the index #

After an import, a database move, or when results are wrong:

php bin/php/updatesearchindex.php --clean --allow-root-user

--clean empties the index first. Run it as the site's user, on a large site in a quiet hour. With eZ Find, reindex with the extension's own tools.

Choosing for a project #

Related pages #

• Searching: the editor's view
• Search statistics
• Search course for trainers
• Free documentation: this page is free to copy, translate and improve under the GNU FDL 1.3.