A developer implemented a custom data table in a Lightning web component with filter functionality.

When a data table experiences slow performance during filtering, the root cause is often "non-selective" queries that result in full table scans. In Salesforce, a query is selective when it uses an indexed field in the WHERE clause and filters down to a small percentage of the total records (typically 10% of the first million).

Option D is the correct strategy. By ensuring that the filter fields (such as a Status or Category) are either standard indexed fields or custom fields marked as "External ID" or "Unique" (which creates a custom index), the Query Optimizer can quickly locate the relevant rows. This significantly reduces the database processing time and the "Time to First Byte" (TTFB) returned to the Lightning component.

Option C might seem attractive for small datasets, but it is not scalable; if the org has 50,000 records, returning all of them to the client will crash the browser's heap memory. Option B (setStorable) is an Aura-specific concept and is replaced by @AuraEnabled(cacheable=true) in LWC, but even caching doesn't help if the initial query itself is slow. Option A (SOSL) is for text searching across multiple objects and is generally less efficient than a targeted SOQL query for specific field filtering. Using a selective SOQL query ensures the component remains fast as data volume grows.