Elasticsearch能否为特定字段搜索设固定分数及按字段匹配次数累加分数？

Absolutely! Let's break down both of your questions with practical, easy-to-follow examples to show you exactly how to achieve this in Elasticsearch.

Yes, you can absolutely set fixed scores for searches targeting specific fields. Two common tools for this are the constant_score query and the function_score query, depending on your needs:

• Using constant_score for straightforward fixed scoring
  If you want every document matching a specific field to get the exact same score (ignoring Elasticsearch's default TF/IDF logic), constant_score is your best bet. It’s ideal for exact matches or filter-based scoring.

  Example: Give all documents where the title field matches "elasticsearch" a fixed score of 5.0:

  GET /your_index/_search
  {
    "query": {
      "constant_score": {
        "filter": {
          "match": {
            "title": "elasticsearch"
          }
        },
        "boost": 5.0
      }
    }
  }
  

• Using function_score for flexible combined scoring
  If you need to mix fixed field scoring with other query logic, function_score lets you add fixed weights to matches on specific fields while keeping a base score from another query.

  Example: Give a base score from matching "distributed systems" in content, then add 3.0 to the score if the tags field matches "database":

  GET /your_index/_search
  {
    "query": {
      "function_score": {
        "query": {
          "match": {
            "content": "distributed systems"
          }
        },
        "functions": [
          {
            "filter": {
              "match": {
                "tags": "database"
              }
            },
            "weight": 3.0
          }
        ],
        "boost_mode": "sum"
      }
    }
  }
  

You can definitely set the score to equal the number of fields that match your query. Here are two simple methods:

• Method 1: Using a bool query with should clauses
  Each should clause targets a different field, and since each clause has a default boost of 1.0, the total score will automatically be the number of matching fields. Just set minimum_should_match to control how many fields need to match for a document to be included.

  Example: Score documents based on how many of title, content, or tags match "elasticsearch":

  GET /your_index/_search
  {
    "query": {
      "bool": {
        "should": [
          { "match": { "title": "elasticsearch" } },
          { "match": { "content": "elasticsearch" } },
          { "match": { "tags": "elasticsearch" } }
        ],
        "minimum_should_match": 1,
        "boost": 1.0
      }
    }
  }
  

  If a document matches both title and tags, its score will be 2.0; if it only matches content, the score is 1.0, and so on.

• Method 2: Using script_score for custom counting
  For more control (like handling partial matches or custom field checks), use a script to count matching fields directly and set the score to that count.

  Example: Count how many fields contain the query term and use that as the score:

  GET /your_index/_search
  {
    "query": {
      "function_score": {
        "query": {
          "multi_match": {
            "query": "elasticsearch",
            "fields": ["title", "content", "tags"]
          }
        },
        "script_score": {
          "script": {
            "source": """
              int matchCount = 0;
              if (doc['title'].value.contains(params.queryTerm)) matchCount++;
              if (doc['content'].value.contains(params.queryTerm)) matchCount++;
              if (doc['tags'].value.contains(params.queryTerm)) matchCount++;
              return matchCount;
            """,
            "params": {
              "queryTerm": "elasticsearch"
            }
          }
        }
      }
    }
  }
  

内容的提问来源于stack exchange，提问作者Karan Jilka