【ES】 语法高亮

三种语法高亮方式

参照B站课程写的

1. plain highlight：默认方式，底层lucene highlight

get /mail/_search
{
  "query": {
    "match": {
      "receiver": "陈一"
    }
  },
  "highlight": {
    "fields": {
      "receiver": {}
    }
  }
}

查询结果最后会有 em标记

"highlight": {
  "receiver": [
    "\"<em>陈一</em>清\" <ds.zhhsl@cn.gree.com>"
  ]
}

2. posting highlight：设置索引的 index_option=offset

优点

1. 性能比plain highlight高，因为不需要重新对高亮文本进行分词
2. 对磁盘的消耗少
3. 将文本切割为句子，并对句子进行高亮，效果更好

索引建立示例

put /mail
{
  "mappings": {
    "properties": {
      "content": {
        "type": "text",
        "analyzer": "standard_multilanguage",
        "index_options": "offsets"
      }
    }
  }
}

查询高亮语法与plain highlight一样，只是性能、效果略好

3. fast vector highlight

优点

1. 对大field（大于1mb）性能更高

索引建立示例

put /mail
{
  "mappings": {
    "properties": {
      "content": {
        "type": "text",
        "analyzer": "standard_multilanguage",
        "term_vector": "with_positions_offsets"
      }
    }
  }
}

查询高亮语法与plain highlight一样

---

强制使用某种高亮方式

get /mail/_search
{
  "query": {
    "match": {
      "receiver": "陈一"
    }
  },
  "highlight": {
    "fields": {
      "receiver": {
      	  "type":"plain"
      }
    }
  }
}