docvalues和Fieldcache

Fieldcache:

 docID->document->fieldvalue

不管是聚類排序關聯等，首先都須要得到文檔中某個字段的值，經過docID去得到整個document，而後再去得到字段值，term轉換獲得最終值，FieldCache一開始就緩存了全部文檔的某個特定域(全部數值類型以及不分詞的stringField)的值到內存，便於隨機存取該域值！

Fieldcache實現過程：

http://moshalanye.iteye.com/blog/281379

  缺點：

1. 常駐內存，大小是全部文檔個數特定域類型大小

2. 初始加載過程耗時，須要遍歷倒排索引及類型轉換

 

Docvalues:

docID->fieldvalue

  建索引時，創建了document到field value的面向列的正排索引數據結構，直接經過已知的docID定位到字段值，從而無需加載document，亦不須要term轉換，遍歷term找尋doc等的過程

   優勢：大約節省三分之一的內存!

   缺點：因爲是硬盤讀取，而非內存模式，對於大批量的使用下，優點明顯，速度更優；小量狀況下沒有內存快！整體會慢15-20%

 

20 February 2015 - Apache Lucene 5.0.0 and Apache Solr 5.0.0 Available

http://lucene.apache.org/ 

FieldCache is gone (moved to a dedicated UninvertingReader in the misc module). This means when you intend to sort on a field, you should index that field using doc values, which is much faster and less heap consuming than FieldCache.

LUCENE-5666：Change uninverted access (sorting, faceting, grouping, etc) to use the DocValues API instead of FieldCache

 

Es中

https://www.elastic.co/guide/en/elasticsearch/guide/current/doc-values.html 

Sorl中

http://wiki.apache.org/solr/DocValues?cm_mc_uid=56088888487714180880058&cm_mc_sid_50200000=1448507379 

https://cwiki.apache.org/confluence/display/solr/DocValues