基于AND运算符的多词Multi Match Query多字段查询实现咨询

Got it, let's build the Elasticsearch query you need for that Udemy-like top bar search with multi-keyword support. Your requirement calls for a Multi Match Query with AND operator across your specified fields, including the nested designers field. Here's a complete, tailored solution:

Final Query DSL #

POST /your_index/_search
{
  "query": {
    "bool": {
      "should": [
        // Match flat fields: category (keyword) and story (text)
        {
          "multi_match": {
            "query": "your_user_search_input",
            "fields": ["category", "story"],
            "operator": "AND",
            "type": "cross_fields",
            "tie_breaker": 0.3
          }
        },
        // Match nested designers.name (keyword)
        {
          "nested": {
            "path": "designers",
            "query": {
              "multi_match": {
                "query": "your_user_search_input",
                "fields": ["designers.name"],
                "operator": "AND"
              }
            },
            "inner_hits": {} // Optional: return matched nested designer objects
          }
        }
      ],
      "minimum_should_match": 1 // Ensure at least one field group matches all keywords
    }
  }
}

Key Breakdown & Why This Works #

Let's walk through the important parts to make sure it aligns with your mappings and use case:

1. Bool Should + Minimum Should Match
   We use should to let the query match either the flat field group or the nested designer field group, then set minimum_should_match: 1 to ensure at least one of these groups satisfies the AND keyword condition. This way, users' multi-keyword input can match across any of your target fields while enforcing all keywords are present in the matched group.

2. Multi Match for Flat Fields

   • operator: "AND": Ensures every keyword in the user's input is found in either category or story (thanks to the cross_fields type, which treats multiple fields as a single combined pool for matching).
   • type: "cross_fields": Perfect for this scenario—it lets keywords split across different fields count as a match (e.g., "java" in category and "beginner" in story still qualifies, whereas best_fields would require both terms in the same field).
   • tie_breaker: Balances scoring between exact keyword matches (like category) and full-text matches (like story) so results feel intuitive.

3. Nested Query for Designers
   Since designers is a nested type, we can't query it directly with a top-level multi-match. The nested clause targets the designers.name keyword field, applying the same AND operator to ensure all user keywords match a designer's name. The optional inner_hits returns which specific designer matched, which is great for displaying context in search results.

Quick Adjustments for Your Needs #

• If you want to require matches in both flat fields and nested designers, swap should for must in the bool clause.
• For case-insensitive matching on keyword fields (e.g., "web dev" matching "Web Dev"), add a lowercase normalizer to your category and designers.name mappings.

This setup gives you the exact Udemy-style search experience where users can type multiple keywords and get results that include all terms across your target fields.

内容的提问来源于stack exchange，提问作者fullmetal