Why analyser dont work with suggester?

Elastic Stack Elasticsearch
1

I'm trying to work on a basic suggester that if someone write "bl" the suggester would propose "blue corola" and "toyota blue". But at the moment its only return "blue corola". It's like the analyser dont work.
Can you help me to find what I am missing

PUT /my-index-000001
{
  "settings": {
    "analysis": {
      "analyzer": {
        "custom_analyzer": {
          "type": "custom",
          "tokenizer": "whitespace",
          "filter": ["lowercase"]
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "color": {
        "type": "completion",
        "analyzer": "custom_analyzer",
        "search_analyzer": "standard"
      }
    }
  }
}

POST /my-index-000001/_doc/1
{
  "color": ["toyota blue"]
}

POST /my-index-000001/_doc/2
{
  "color": ["toyota blue"]
}

POST /my-index-000001/_doc/3
{
  "color": ["blue corola"]
}


POST /my-index-000001/_search
{
  "suggest": {
    "my-suggestion": {
      "text": "bl",
      "completion": {
        "field": "color",
        "skip_duplicates": "false"
      }
    }
  }
}
2

Hi @Adrien_L Welcome!

Try to separate the terms in indexing.

Doc: Suggesters | Elasticsearch Guide [7.17] | Elastic

POST /my-index-000001/_doc/1
{
  "color": ["toyota", "blue"]
}

POST /my-index-000001/_doc/2
{
  "color": ["toyota", "blue"]
}

POST /my-index-000001/_doc/3
{
  "color": ["blue", "corola"]
}
3

Thank you @RabBit_BR !
Okay so yes, it works with your manual splitting technique but ideally, I would like to use Elasticsearch tokenizer and not do it manually. As you can see I set:
"tokenizer": "whitespace"
so I' m guessing it already created an index with each word tokenize and ready to be returned for my suggestion no?