{"section":"known-issues","requestedLocale":"en","requestedSlug":"sampling-for-search-filters-unexpectedly-aggressive-omitting-filter-values","locale":"en","slug":"sampling-for-search-filters-unexpectedly-aggressive-omitting-filter-values","path":"docs/en/known-issues/Intelligent Search/sampling-for-search-filters-unexpectedly-aggressive-omitting-filter-values.md","branch":"main","content":"## Summary\n\n\nWhen a product search finds more than 30.000 results, for efficiency purposes, it is defined that attribute filters are computed on top of only the first 30.000 most relevant products (following the store's relevance settings) inside the specific search.\n\nDue to a problem, filters are actually being computed over products that are not from the specific search, reducing the universe of filters that should be offered because they are from unrelated products.\n\n\n#### Simulation\n\n\nTo simulate the scenario, consider something like a store with 500k products and a specific category with 35k products.\n\nWhile browsing this category, since the sampling is expected to use the 30k more relevant products, filters should be close enough to all the values the category can offer. However, with this problem, it may offer fewer filters than expected because the sampling was filled with products from other categories whose filters won't be considered.\n\n\n#### Workaround\n\n\nN/A"}