Skip to content

Query Performance

When to Use

When optimizing search query speed and reducing server load.

Decision: Performance Strategies

Strategy Impact Applies To
Whole words only matching Significant speedup Database backend
Skip result count query Major speedup on MySQL InnoDB Database backend
Mini pager Avoids COUNT query All backends
Skip entity loads (Solr index-only) Eliminates DB queries Solr backend
Highlight processor skip 10x latency reduction All backends
Cache warming Faster repeated queries All backends

Pattern: MySQL InnoDB COUNT Problem

MySQL InnoDB COUNT queries are extremely slow on large tables. Two mitigations:

  1. Use Mini pager in Views instead of Full pager — Mini pager doesn't need total count
  2. Enable "Skip result count query" in Views query settings

Pattern: Highlight Performance

The Highlight processor can add 10x latency on complex queries. Options: - Disable globally if excerpts aren't needed - Skip for specific pages via query tag: 

$query->addTag('search_api_skip_processor_highlight');

Pattern: Cache Warming

For high-traffic search pages: 1. Identify popular search terms from analytics 2. Create a cron job that pre-executes popular queries 3. This warms Drupal's render cache and the backend's query cache

For Solr: configure autowarming in solrconfig.xml for filter cache, query result cache, and document cache.

Common Mistakes

  • Using AJAX on search results Views — Breaks unique URLs, harms UX and analytics. Does not improve performance.
  • Partial matching on DB backend — "Partial matching" is much slower than "Whole words only." Only use partial if required.

See Also