白话Elasticsearch10-深度探秘搜索技术之基于dis_max实现best fields策略进行多字段搜索
文章目录
- 概述
- TF/IDF
- 链接
- 示例
- DSL
- 普通查询
- dis_max 查询
- best fields策略-dis_max
概述
继续跟中华石杉老师学习ES，第十篇
课程地址： https://www.roncoo.com/view/55
TF/IDF
Apache Lucene默认评分机制
-
TF (Term Frequency): 基于词项(term vector), 用来表示一个词项在某个文档中出现了多少次。
词频越高，文档得分越高 -
IDF (Inveres Dcoument Frequency): 基于词项（term vector）,用来告诉评分公式该词有多美的汉奸。
逆文档频率越高，词项就越罕见。 评分公式利用该因子为包含罕见词项的文档加权。
term vector : 词项向量是一种针对每个文档的微型倒排索引。词项向量的每个维由词项和出现频率结对组成,还可以包含词项的位置信息。 Lucene 和 ES都默认禁用词项向量索引，如果实现某些功能比如高亮显示等需要开启该选项 。
链接
官方指导： https://www.elastic.co/guide/en/elasticsearch/guide/current/_tuning_best_fields_queries.html
https://www.elastic.co/guide/en/elasticsearch/reference/7.2/query-dsl-dis-max-query.html
数据量少的时候，dis_max不生效的问题： https://stackoverflow.com/questions/38065692/dis-max-query-isnt-looking-for-the-best-matching-clause
其他博主写的相关文章：
https://blog.csdn.net/dm_vincent/article/details/41820537
示例
ES版本 6.4.1
为了演示效果，我们把之前的forum索引删除了重建一下，
DSL如下
DSL
DELETE /forum
PUT /forum
{ "settings" : { "number_of_shards" : 1 }}
POST /forum/article/_bulk
{ "index": { "_id": 1 }}
{ "articleID" : "XHDK-A-1293-#fJ3", "userID" : 1, "hidden": false, "postDate": "2017-01-01" }
{ "index": { "_id": 2 }}
{ "articleID" : "KDKE-B-9947-#kL5", "userID" : 1, "hidden": false, "postDate": "2017-01-02" }
{ "index": { "_id": 3 }}
{ "articleID" : "JODL-X-1937-#pV7", "userID" : 2, "hidden": false, "postDate": "2017-01-01" }
{ "index": { "_id": 4 }}
{ "articleID" : "QQPX-R-3956-#aD8", "userID" : 2, "hidden": true, "postDate": "2017-01-02" }
POST /forum/article/_bulk
{"update":{"_id":"1"}}
{"doc":{"tag":["java","hadoop"]}}
{"update":{"_id":"2"}}
{"doc":{"tag":["java"]}}
{"update":{"_id":"3"}}
{"doc":{"tag":["hadoop"]}}
{"update":{"_id":"4"}}
{"doc":{"tag":["java","elasticsearch"]}}
POST /forum/article/_bulk
{"update":{"_id":"1"}}
{"doc":{"tag_cnt":2}}
{"update":{"_id":"2"}}
{"doc":{"tag_cnt":1}}
{"update":{"_id":"3"}}
{"doc":{"tag_cnt":1}}
{"update":{"_id":"4"}}
{"doc":{"tag_cnt":2}}
POST /forum/article/_bulk
{"update":{"_id":"1"}}
{"doc":{"view_cnt":30}}
{"update":{"_id":"2"}}
{"doc":{"view_cnt":50}}
{"update":{"_id":"3"}}
{"doc":{"view_cnt":100}}
{"update":{"_id":"4"}}
{"doc":{"view_cnt":80}}
POST /forum/article/_bulk
{"index":{"_id":5}}
{"articleID":"DHJK-B-1395-#Ky5","userID":3,"hidden":false,"postDate":"2019-06-01","tag":["elasticsearch"],"tag_cnt":1,"view_cnt":10}
POST /forum/article/_bulk
{"update":{"_id":"5"}}
{"doc":{"postDate":"2019-05-01"}}
POST /forum/article/_bulk
{"update":{"_id":"1"}}
{"doc":{"title":"this is java and elasticsearch blog"}}
{"update":{"_id":"2"}}
{"doc":{"title":"this is java blog"}}
{"update":{"_id":"3"}}
{"doc":{"title":"this is elasticsearch blog"}}
{"update":{"_id":"4"}}
{"doc":{"title":"this is java, elasticsearch, hadoop blog"}}
{"update":{"_id":"5"}}
{"doc":{"title":"this is spark blog"}}
POST /forum/article/_bulk
{"update":{"_id":"1"}}
{"doc":{"content":"i like to write best elasticsearch article"}}
{"update":{"_id":"2"}}
{"doc":{"content":"i think java is the best programming language"}}
{"update":{"_id":"3"}}
{"doc":{"content":"i am only an elasticsearch beginner"}}
{"update":{"_id":"4"}}
{"doc":{"content":"elasticsearch and hadoop are all very good solution, i am a beginner"}}
{"update":{"_id":"5"}}
{"doc":{"content":"spark is best big data solution based on scala ,an programming language similar to java"}}
至此，数据构造完成 ，下面来看下dis_max是如何作用的吧
GET /forum/article/_search
数据如下：
{
"took": 0,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"skipped": 0,
"failed": 0
},
"hits": {
"total": 5,
"max_score": 1,
"hits": [
{
"_index": "forum",
"_type": "article",
"_id": "1",
"_score": 1,
"_source": {
"articleID": "XHDK-A-1293-#fJ3",
"userID": 1,
"hidden": false,
"postDate": "2017-01-01",
"tag": [
"java",
"hadoop"
],
"tag_cnt": 2,
"view_cnt": 30,
"title": "this is java and elasticsearch blog",
"content": "i like to write best elasticsearch article"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "2",
"_score": 1,
"_source": {
"articleID": "KDKE-B-9947-#kL5",
"userID": 1,
"hidden": false,
"postDate": "2017-01-02",
"tag": [
"java"
],
"tag_cnt": 1,
"view_cnt": 50,
"title": "this is java blog",
"content": "i think java is the best programming language"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "3",
"_score": 1,
"_source": {
"articleID": "JODL-X-1937-#pV7",
"userID": 2,
"hidden": false,
"postDate": "2017-01-01",
"tag": [
"hadoop"
],
"tag_cnt": 1,
"view_cnt": 100,
"title": "this is elasticsearch blog",
"content": "i am only an elasticsearch beginner"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "4",
"_score": 1,
"_source": {
"articleID": "QQPX-R-3956-#aD8",
"userID": 2,
"hidden": true,
"postDate": "2017-01-02",
"tag": [
"java",
"elasticsearch"
],
"tag_cnt": 2,
"view_cnt": 80,
"title": "this is java, elasticsearch, hadoop blog",
"content": "elasticsearch and hadoop are all very good solution, i am a beginner"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "5",
"_score": 1,
"_source": {
"articleID": "DHJK-B-1395-#Ky5",
"userID": 3,
"hidden": false,
"postDate": "2019-05-01",
"tag": [
"elasticsearch"
],
"tag_cnt": 1,
"view_cnt": 10,
"title": "this is spark blog",
"content": "spark is best big data solution based on scala ,an programming language similar to java"
}
}
]
}
}
普通查询
先看下普通的DSL
GET /forum/article/_search
{
"query": {
"bool": {
"should": [
{
"match": {
"title": "java solution"
}
},
{
"match": {
"content": "java solution"
}
}
],
"minimum_should_match": 1
}
}
}
返回：
{
"took": 1,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"skipped": 0,
"failed": 0
},
"hits": {
"total": 4,
"max_score": 1.5179626,
"hits": [
{
"_index": "forum",
"_type": "article",
"_id": "2",
"_score": 1.5179626,
"_source": {
"articleID": "KDKE-B-9947-#kL5",
"userID": 1,
"hidden": false,
"postDate": "2017-01-02",
"tag": [
"java"
],
"tag_cnt": 1,
"view_cnt": 50,
"title": "this is java blog",
"content": "i think java is the best programming language"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "5",
"_score": 1.4233948,
"_source": {
"articleID": "DHJK-B-1395-#Ky5",
"userID": 3,
"hidden": false,
"postDate": "2019-05-01",
"tag": [
"elasticsearch"
],
"tag_cnt": 1,
"view_cnt": 10,
"title": "this is spark blog",
"content": "spark is best big data solution based on scala ,an programming language similar to java"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "4",
"_score": 1.2832261,
"_source": {
"articleID": "QQPX-R-3956-#aD8",
"userID": 2,
"hidden": true,
"postDate": "2017-01-02",
"tag": [
"java",
"elasticsearch"
],
"tag_cnt": 2,
"view_cnt": 80,
"title": "this is java, elasticsearch, hadoop blog",
"content": "elasticsearch and hadoop are all very good solution, i am a beginner"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "1",
"_score": 0.4889865,
"_source": {
"articleID": "XHDK-A-1293-#fJ3",
"userID": 1,
"hidden": false,
"postDate": "2017-01-01",
"tag": [
"java",
"hadoop"
],
"tag_cnt": 2,
"view_cnt": 30,
"title": "this is java and elasticsearch blog",
"content": "i like to write best elasticsearch article"
}
}
]
}
}
来分析一下结果
计算每个document的relevance score：每个query的分数，乘以matched query数量，除以总query数量
算一下doc2的分数
{ "match": { "title": "java solution" }}，针对doc2，是有一个分数的
{ "match": { "content": "java solution" }}，针对doc2，也是有一个分数的
假设分数如下 ， 所以是两个分数加起来，比如说，1.1 + 1.2 = 2.3
matched query数量 = 2
总query数量 = 2
2.3 * 2 / 2 = 2.3
算一下doc5的分数
{ "match": { "title": "java solution" }}，针对doc5，是没有分数的
{ "match": { "content": "java solution" }}，针对doc5，是有一个分数的
所以说，只有一个query是有分数的，比如2.3
matched query数量 = 1
总query数量 = 2
2.3 * 1 / 2 = 1.15
doc5的分数 = 1.15 < doc2的分数 = 2.3
id=2的数据排在了前面，其实我们希望id=5的排在前面，毕竟id=5的数据 content字段既有java又有solution. 那看下dis_max吧
dis_max 查询
GET /forum/article/_search
{
"query": {
"dis_max": {
"queries": [
{
"match": {
"title": "java solution"
}
},
{
"match": {
"content": "java solution"
}
}
]
}
}
}
返回
{
"took": 0,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"skipped": 0,
"failed": 0
},
"hits": {
"total": 4,
"max_score": 1.4233948,
"hits": [
{
"_index": "forum",
"_type": "article",
"_id": "5",
"_score": 1.4233948,
"_source": {
"articleID": "DHJK-B-1395-#Ky5",
"userID": 3,
"hidden": false,
"postDate": "2019-05-01",
"tag": [
"elasticsearch"
],
"tag_cnt": 1,
"view_cnt": 10,
"title": "this is spark blog",
"content": "spark is best big data solution based on scala ,an programming language similar to java"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "2",
"_score": 0.93952733,
"_source": {
"articleID": "KDKE-B-9947-#kL5",
"userID": 1,
"hidden": false,
"postDate": "2017-01-02",
"tag": [
"java"
],
"tag_cnt": 1,
"view_cnt": 50,
"title": "this is java blog",
"content": "i think java is the best programming language"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "4",
"_score": 0.79423964,
"_source": {
"articleID": "QQPX-R-3956-#aD8",
"userID": 2,
"hidden": true,
"postDate": "2017-01-02",
"tag": [
"java",
"elasticsearch"
],
"tag_cnt": 2,
"view_cnt": 80,
"title": "this is java, elasticsearch, hadoop blog",
"content": "elasticsearch and hadoop are all very good solution, i am a beginner"
}
},
{
"_index": "forum",
"_type": "article",
"_id": "1",
"_score": 0.4889865,
"_source": {
"articleID": "XHDK-A-1293-#fJ3",
"userID": 1,
"hidden": false,
"postDate": "2017-01-01",
"tag": [
"java",
"hadoop"
],
"tag_cnt": 2,
"view_cnt": 30,
"title": "this is java and elasticsearch blog",
"content": "i like to write best elasticsearch article"
}
}
]
}
}
best fields策略-dis_max
best fields策略 : 搜索到的结果，应该是某一个field中匹配到了尽可能多的关键词，被排在前面；而不是尽可能多的field匹配到了少数的关键词，排在了前面.
dis_max语法，直接取多个query中，分数最高的那一个query的分数即可
举个例子
{ "match": { "title": "java solution" }}，针对doc2，是有一个分数的，1.1
{ "match": { "content": "java solution" }}，针对doc2，也是有一个分数的，1.2
取最大分数，1.2
{ "match": { "title": "java solution" }}，针对doc5，是没有分数的
{ "match": { "content": "java solution" }}，针对doc5，是有一个分数的，2.3
取最大分数，2.3
然后doc2的分数 = 1.2 < doc5的分数 = 2.3，所以doc5就可以排在更前面的地方.