Elasticsearch添加索引分析器后无效果，求技术解决方案

Hey there! Let's figure out why your Elasticsearch analyzers aren't working and get this sorted out step by step.

First, the #1 Mistake New ES Users Make #

Elasticsearch doesn't let you modify the analyzer of an existing field after the index is created—even if you close and re-open the index. The analyzer for a field is locked in the moment the field is first mapped. So when you added analyzers post-index-creation, your original title field kept using the initial english analyzer (or whatever default was applied if you didn't set it correctly upfront), and any new analyzers you added didn't affect it.

Step 1: Verify Your Analyzers Are Actually Working #

Before troubleshooting searches, confirm your analyzers are producing the tokens you expect using the _analyze API. This is the fastest way to debug:

Test the built-in english analyzer:

POST /your_test_index/_analyze
{
  "analyzer": "english",
  "text": "movement"
}

You should see a token like move in the response—this is the stemmed form the analyzer generates. If you don't get this, your analyzer isn't configured right (or you're targeting the wrong index).

Test your custom analyzer (e.g., my_lowercaser):

POST /your_test_index/_analyze
{
  "analyzer": "my_lowercaser",
  "text": "movement"
}

Check if the output matches what you expect (e.g., lowercase movement if that's all your custom analyzer does).

Step 2: Fix Existing Fields & Create a Proper Index #

Since you can't change the analyzer on an existing title field, your best bet is to create a new index with the correct analyzer configuration from the start, then re-import your data. Here's a proper example:

PUT /proper_test_index
{
  "settings": {
    "analysis": {
      "analyzer": {
        # Example custom English analyzer (tweak filters as needed)
        "my_custom_english": {
          "tokenizer": "standard",
          "filter": [
            "lowercase",
            "english_stop",
            "english_stemmer"
          ]
        },
        # Your custom lowercase analyzer
        "my_lowercaser": {
          "tokenizer": "standard",
          "filter": ["lowercase"]
        }
      },
      "filter": {
        "english_stop": {
          "type": "stop",
          "stopwords": "_english_"
        },
        "english_stemmer": {
          "type": "stemmer",
          "language": "english"
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "title": {
        "type": "text",
        "analyzer": "my_custom_english" # Bind analyzer to field upfront
      },
      "title2": {
        "type": "text",
        "analyzer": "english" # Use built-in English analyzer
      }
    }
  }
}

Step 3: Effective Query Example #

Once your index is set up correctly, a simple match query will work because ES uses the same analyzer for indexing and searching by default:

POST /proper_test_index/_search
{
  "query": {
    "match": {
      "title": "moving"
    }
  }
}

This should return documents with "movement" in the title field, since both "moving" and "movement" get stemmed to move by the English analyzer.

If you need to explicitly specify the analyzer for searching (for debugging), use this:

POST /proper_test_index/_search
{
  "query": {
    "match": {
      "title": {
        "query": "moving",
        "analyzer": "my_custom_english"
      }
    }
  }
}

Additional Debugging Tips #

• Check field mappings: Confirm your fields are actually using the right analyzer with the _mapping API:

  GET /proper_test_index/_mapping
  

• Inspect indexed terms: Use the _termvectors API to see exactly what tokens were stored for a document:

  GET /proper_test_index/_doc/your_document_id/_termvectors?fields=title
  

  Look for the stemmed token (like move) in the response—if it's missing, your analyzer didn't process the field correctly during indexing.

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