关于 keep_first_letter 和 keep_joined_full_pinyin 在 suggest completion 的一些疑惑

[zhengpq]

在实际使用中发现两个有趣的现象：

• 输入 first letter 的情况没有办法匹配到结果
• 输入 full pinyin 的情况下，在输入过程中如果某个字的拼音没有输全的话也会匹配不到

具体见下方：

索引

{
  "settings": {
    "analysis": {
      "analyzer": {
        "pinyin_ik": {
          "tokenizer": "ik_smart",
          "filter": [
            "py"
          ]
        },
        "pinyin_keyword": {
          "tokenizer": "keyword",
          "filter": [
            "py"
          ]
        },
        "pinyin_keyword_full": {
          "tokenizer": "keyword",
          "filter": [
            "py_full"
          ]
        }
      },
      "filter": {
        "py": {
          "type": "pinyin",
          "keep_original": true,
          "keep_first_letter": true,
          "keep_full_pinyin": true,
          "keep_joined_full_pinyin": true,
          "keep_none_chinese_in_joined_full_pinyin": true,
          "limit_first_letter_length": 16
        },
        "py_full": {
          "type": "pinyin",
          "keep_original": true,
          "keep_first_letter": false,
          "keep_full_pinyin": true,
          "keep_joined_full_pinyin": true,
          "keep_none_chinese_in_joined_full_pinyin": true,
          "limit_first_letter_length": 16
        }
      }
    }
  },
  "mappings": {
    "properties": {
      "page_id": {
        "type": "integer"
      },
      "page_title": {
        "type": "text",
        "analyzer": "pinyin",
        "search_analyzer": "ik_smart",
        "fields": {
          "suggest": {
            "type": "completion",
            "analyzer": "pinyin_keyword",
            "search_analyzer": "pinyin_keyword_full"
          }
        }
      },
    }
  }
}

### 1、首字母没有办法匹配到

文章标题分析结果

POST wxad-page-new/_analyze
{
  "analyzer": "pinyin_keyword",
  "text": "公众号流量主基础介绍"
}

// 结果，部分无关的 token 删除

{
  "tokens" : [
    {
      "token" : "gzhllzjcjs",
      "start_offset" : 0,
      "end_offset" : 10,
      "type" : "word",
      "position" : 9
    }
  ]
}

搜索及结果

POST wxad-page-new/_search?pretty
{
  "_source": {
    "includes": [
      "page_title"
    ]
  },
  "suggest": {
    "title-suggest": {
      "prefix": "gzhl",
      "completion": {
        "field": "page_title.suggest",
        "size": 100
      }
    }
  }
}

// 结果

{
  "took" : 0,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 0,
      "relation" : "eq"
    },
    "max_score" : null,
    "hits" : [ ]
  },
  "suggest" : {
    "title-suggest" : [
      {
        "text" : "gzhl",
        "offset" : 0,
        "length" : 4,
        "options" : [ ]
      }
    ]
  }
}

### 2、输入全部拼音的问题

文章标题分析结果

POST wxad-page-new/_analyze
{
  "analyzer": "pinyin_keyword",
  "text": "公众号流量主基础介绍"
}

// 结果，部分无关的 token 删除

{
  "tokens" : [
   {
      "token" : "gongzhonghaoliuliangzhujichujieshao",
      "start_offset" : 0,
      "end_offset" : 10,
      "type" : "word",
      "position" : 0
    }
  ]
}

搜索及结果——能够匹配到的情况

POST wxad-page-new/_search?pretty
{
  "_source": {
    "includes": [
      "page_title"
    ]
  },
  "suggest": {
    "title-suggest": {
      "prefix": "gongzhong",
      "completion": {
        "field": "page_title.suggest",
        "size": 100
      }
    }
  }
}

// 结果

{
  "took" : 4,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 0,
      "relation" : "eq"
    },
    "max_score" : null,
    "hits" : [ ]
  },
  "suggest" : {
    "title-suggest" : [
      {
        "text" : "gongzhong",
        "offset" : 0,
        "length" : 9,
        "options" : [
        {
            "text" : "公众号流量主基础介绍",
            "_index" : "wxad-page-new",
            "_type" : "_doc",
            "_id" : "259",
            "_score" : 1.0,
            "_source" : {
              "page_title" : "公众号流量主基础介绍"
            }
          }
        ]
      }
    ]
  }
}

搜索及结果——匹配不到的情况

POST wxad-page-new/_search?pretty
{
  "_source": {
    "includes": [
      "page_title"
    ]
  },
  "suggest": {
    "title-suggest": {
      "prefix": "gongzhon",
      "completion": {
        "field": "page_title.suggest",
        "size": 100
      }
    }
  }
}

// 结果

{
  "took" : 0,
  "timed_out" : false,
  "_shards" : {
    "total" : 1,
    "successful" : 1,
    "skipped" : 0,
    "failed" : 0
  },
  "hits" : {
    "total" : {
      "value" : 0,
      "relation" : "eq"
    },
    "max_score" : null,
    "hits" : [ ]
  },
  "suggest" : {
    "title-suggest" : [
      {
        "text" : "gongzhon",
        "offset" : 0,
        "length" : 8,
        "options" : [ ]
      }
    ]
  }
}

差了一个 g 就匹配不到，让我跟感到很费解。

由于是新手，不知道这个是 pinyin 这个插件的特性还是 es 本身的一些特性，如果有哪位小伙伴知道这个问题的解决方案，麻烦告知一下，感谢！