Elasticsearch Field Options Norms

Elasticsearch 定義字段時Norms選項的做用

本文介紹ElasticSearch中2種字段(text 和 keyword)的Norms參數做用。html

建立ES索引時,通常指定2種配置信息:settings、mappings。settings 與數據存儲有關(幾個分片、幾個副本);而mappings 是數據模型,相似於MySQL中的表結構定義。在Mapping信息中指定每一個字段的類型,ElasticSearch支持多種類型的字段(field datatypes),好比String、Numeric、Date…其中String又細分紅爲種:keyword 和 text。在建立索引時,須要定義字段併爲每一個字段指定類型,示例以下:java

PUT my_index
{
  "settings": {
    "number_of_shards": 1,
    "number_of_replicas": 0
  },
  "mappings": {
    "_doc": {
      "_source": {
        "enabled": true
      },
      "properties": {
        "title": {
          "type": "text",
          "norms": false
        },
        "overview": {
          "type": "text",
          "norms": true
        },
        "body": {
          "type": "text"
        },
        "author": {
          "type": "keyword",
          "norms": true
        },
        "chapters": {
          "type": "keyword",
          "norms": false
        },
        "email": {
          "type": "keyword"
        }
      }
    }
  }
}

my_index 索引的 title 字段類型是 text,而 author 字段類型是 keyword。算法

對於 text 類型的字段而言,默認開啓了norms,而 keyword 類型的字段則默認關閉了normsapp

Whether field-length should be taken into account when scoring queries. Accepts true(text filed datatype) or false(keyword filed datatype)elasticsearch

爲何 keyword 類型的字段默認關閉 norms 呢?keyword 類型的string 可理解爲:Do index the field, but don't analyze the string value,也即:keyword 類型的字段是不會被Analyzer "分析成" 一個個的term的,它是一個single-token fields,所以也就不須要字段長度(fieldNorm)、tfNorm(term frequency Norm)這些歸一化因子了。而 text 類型的字段會被分析器(Analyzer)分析,生成若干個terms,兩個 text 類型的字段,一個可能有不少term(好比文章的正文),另外一個只有不多的term(好比文章的標題),在多字段查詢時,就須要長度歸一化,這就是爲何 text 類型字段默認開啓 norms 選項的緣由吧。另外,對於Lucene經常使用的2種評分算法:tf-idf 和 bm25,tf-idf 就傾向於給長度較小的字段打高分,爲何呢?Lucene 的類似度評分公式,主要由三部分組成:IDF score,TF score 還有 fieldNorms。就TF-IDF評分公式而言,IDF score 是log(numDocs/(docFreq+1)),TF score 是 sqrt(tf),fieldNorms 是 1/sqrt(length),所以:文檔長度越短,fieldNorms越大,評分越高,這也是爲何TF-IDF嚴重偏向於給短文本打高分的緣由。ide

norms 做用是什麼?

norms 是一個用來計算文檔/字段得分(Score)的"調節因子"。TF-IDF、BM25算法計算文檔得分時都用到了norms參數,具體可參考這篇文章中的Lucene文檔得分計算公式。ui

ElasticSearch中的一篇文檔(Document),裏面有多個字段。查詢解析器(QueryParser)將用戶輸入的查詢字符串解析成Terms ,在多字段搜索中,每一個 Term 會去匹配各個字段,爲每一個字段計算一個得分,各個字段的得分通過某種方式(以詞爲中心的搜索 vs 以字段爲中心的搜索)組合起來,最終獲得一篇文檔的得分。this

ES官方文檔關於Norms解釋:code

Norms store various normalization factors that are later used at query time in order to compute the score of a document relatively to a query.orm

這裏的 normalization factors 用於查詢計算文檔得分時進行 boosting。好比根據BM25算法給出的公式(freq*(k1+1))/(freq+k1*(1-b+b*fieldLength/avgFieldLength))計算文檔得分時,其中的fieldLength/avgFieldLength就是 normalization factors。

norms 的代價

開啓norms以後,每篇文檔的每一個字段須要一個字節存儲norms。對於 text 類型的字段而言是默認開啓norms的,所以對於不須要評分的 text 類型的字段,能夠禁用norms,這算是一個調優勢吧。

Although useful for scoring, norms also require quite a lot of disk (typically in the order of one byte per document per field in your index, even for documents that don’t have this specific field). As a consequence, if you don’t need scoring on a specific field, you should disable norms on that field

norms 因子屬於 Index-time boosting一部分,也即:在索引文檔(寫入文檔)的時候,就已經將全部boosting因子存儲起來,在查詢時從內存中讀取,參與得分計算。參考《Lucene in action》中一段話:

During indexing, all sources of index-time boosts are combined into a single floating point number for each indexed field in the document. The document may have its own boost; each field may have a boost; and Lucene computes an automatic boost based on the number of tokens in the field (shorter fields have a higher boost). These boosts are combined and then compactly encoded (quantized) into a single byte, which is stored per field per document. During searching, norms for any field being searched are loaded into memory, decoded back into a floating-point number, and used when computing the relevance score.

另外一種類型的 boosting 是search time boosting,在查詢語句中指定boosting因子,而後動態計算出文檔得分,具體可參考:《relevant search with applications for solr and elasticsearch》,本文再也不詳述。可是值得注意的是:目前的ES版本已經再也不推薦使用index time boosting了,而是推薦使用 search time boosting。ES官方文檔給出的理由以下:

  • 在索引文檔時存儲的boosting因子(開啓 norms 選項),一經存儲,就沒法改變。要想改變,只能reindex索引
  • search time boosting 的效果和 index time boosting是同樣的,而且search time boosting可以動態指定boosting因子(但計算文檔得分時更消耗CPU吧),靈活性更大。而index time boosting須要額外的存儲空間
  • index time boosting因子存儲在norms字段,它影響了 field length normalization,從而致使文檔類似度計算結果不太準確(lower quality relevance calculations)

附:my_index索引的mapping 信息:

GET my_index/_mapping

{
  "my_index": {
    "mappings": {
      "_doc": {
        "properties": {
          "author": {
            "type": "keyword",
            "norms": true
          },
          "body": {
            "type": "text"
          },
          "chapters": {
            "type": "keyword"
          },
          "email": {
            "type": "keyword"
          },
          "overview": {
            "type": "text"
          },
          "title": {
            "type": "text",
            "norms": false
          }
        }
      }
    }
  }
}

原文:http://www.javashuo.com/article/p-dnaymfdf-bo.html

相關文章
相關標籤/搜索