Elasticsearch：Completion Suggester与Whitespace Analyzer兼容问题咨询

Hey Jimmy, let's break down why your "ela" prefix isn't returning the expected result, and confirm that Whitespace Analyzer does work with Completion Suggester—you just need to make sure everything is set up correctly!

Why Your Current Setup Isn't Working #

The core issue here is almost certainly that your existing document wasn't re-indexed after you added the custom completion_analyzer to your mapping. Here's what's happening under the hood:

• When you define an analyzer for a completion field, Elasticsearch uses it to split the input string into individual tokens. For your input "Hello elastic search.", the Whitespace Analyzer (plus lowercase filter) generates three tokens: hello, elastic, search..
• Each of these tokens gets its own prefix index (e.g., elastic creates entries for e, el, ela, elas, etc.).
• If you didn't re-index your document after updating the mapping, the suggest field is still using the default behavior: treating the entire input string as a single token. That's why "hel" works (it matches the start of the full string "Hello elastic search."), but "ela" doesn't (it doesn't match the start of the full string).

Step-by-Step Solution #

1. Re-index your documents
   After updating the mapping, you need to re-index existing documents so the completion_analyzer processes the suggest input correctly. You can do this by deleting and re-adding the document, or use the re-index API if you have multiple documents to update.

2. Verify tokenization (optional but helpful)
   You can test how your analyzer processes the input string using the Analyze API to confirm it's generating the right tokens:

   POST _analyze
   {
     "analyzer": "completion_analyzer",
     "text": "Hello elastic search."
   }
   

   This should return the tokens hello, elastic, search.—exactly what we need for prefix matching.

3. Test your search again
   Once the document is re-indexed, your search request with "prefix": "ela" should now match the elastic token's prefix and return the expected document.

Understanding Analyzer's Role in Completion Suggester #

To answer your original question: Yes, analyzers are fully functional with Completion Suggester. Their purpose is to break down the input string into multiple tokens, so each token becomes a separate entry point for prefix matching. This means users can search for prefixes of any word in your input string, not just the first word—exactly what you want when you want "ela" to find your document with "elastic" in it.

If you had manually set input: ["Hello","elastic","search"], you'd achieve a similar result, but using an analyzer automates this process for longer or dynamic input strings (e.g., user-generated content where you don't want to manually split words).

Final Notes #

• Ensure your Elasticsearch version supports the completion field type (it's available in all recent versions, but double-check if you're on a very old release).
• If you want to handle punctuation (like the . at the end of "search."), you can add a trim or pattern_replace filter to your analyzer to remove unwanted characters.

内容的提问来源于stack exchange，提问作者Jimmy