ES DSL搜索 - multi_match、boost和布尔查询
1 介绍
主要介绍索引请求的基础API操作，使用postman进行请求，接口请求的前缀地址统一为elasticsearch 部署IP地址+端口号（例如 http://192.168.51.4:9200 。
统一请求地址：
POST /search_demo/_doc/_search
2 multi_match
multi_match 满足使用 match 在多个字段中进行查询的需求
传递JSON数据
{
"query": {
"multi_match": {
"query": "组合",
"fields": [
"desc","nickname"
]
}
},
"_source": [
"id",
"nickname",
"desc"
]
}
请求结果
{
"took": 3,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"skipped": 0,
"failed": 0
},
"hits": {
"total": {
"value": 1,
"relation": "eq"
},
"max_score": 2.2874916,
"hits": [
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1007",
"_score": 2.2874916,
"_source": {
"nickname": "老男孩",
"id": 1007,
"desc": "确实是个很好的组合，筷子 筷子"
}
}
]
}
}
3 boost
boost 权重，为某个字段设置权重，权重越高，文档相关性得分就越高。通常来说，搜索商品名称要比商品简介的权重要高。
传递JSON数据
{
"query": {
"multi_match": {
"query": "好的",
"fields": [
"desc","nickname*10"
]
}
},
"_source": [
"id",
"nickname",
"desc"
]
}
请求结果
{
"took": 4,
"timed_out": false,
"_shards": {
"total": 1,
"successful": 1,
"skipped": 0,
"failed": 0
},
"hits": {
"total": {
"value": 9,
"relation": "eq"
},
"max_score": 3.1980762,
"hits": [
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1004",
"_score": 3.1980762,
"_source": {
"nickname": "红帽子",
"id": 1004,
"desc": "好的系统必须拥有稳定的系统结构"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1005",
"_score": 3.0979095,
"_source": {
"nickname": "switch游戏机",
"id": 1005,
"desc": "好的游戏，才会有人购买，比如塞尔达"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1003",
"_score": 0.37556386,
"_source": {
"nickname": "涡轮增压",
"id": 1003,
"desc": "极限的速度是需要涡轮增压的"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1012",
"_score": 0.36424035,
"_source": {
"nickname": "youzi",
"id": 1012,
"desc": "永远的神"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1002",
"_score": 0.35254776,
"_source": {
"nickname": "进击的巨人",
"id": 1002,
"desc": "艾伦是会变成真正的巨人的"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1011",
"_score": 0.27665582,
"_source": {
"nickname": "皮特",
"id": 1011,
"desc": "皮特的姓氏好像是彼得"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1007",
"_score": 0.2639615,
"_source": {
"nickname": "老男孩",
"id": 1007,
"desc": "确实是个很好的组合，筷子 筷子"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1009",
"_score": 0.252381,
"_source": {
"nickname": "露西",
"id": 1009,
"desc": "露西是一只很聪明的cat"
}
},
{
"_index": "search_demo",
"_type": "_doc",
"_id": "1001",
"_score": 0.18093815,
"_source": {
"nickname": "飞翔的荷兰号",
"id": 1001,
"desc": "我在p2pi网站解决项目中遇到的问题，学习到了很多知识"
}
}
]
}
}
nickname^10 代表搜索提升10倍的相关性，也就是说用户搜索的时候，其实以这个nickname为主，desc为辅，nickname的匹配相关度高只要提升权重比例就可以。
4 布尔查询
可以实现多重组合查询
- must：查询必须匹配搜索条件，譬如 and
- should：查询匹配满足一个条件，譬如 or
- must_not：不匹配搜索条件，一个都不满足，譬如 not in
单个查询
{
"query": {
"bool": {
"must": [
{
"multi_match": {
"query": "好的",
"fields": ["desc","nickname"]
}
},
{
"term": {
"sex": 0
}
},
{
"term": {
"birthday": "1992-12-24"
}
}
]
}
},
"_source": [
"id",
"sex",
"nickname",
"desc"
]
}
组合查询
{
"query": {
"bool": {
"must": [
{
"match": {
"desc": "好的"
}
},
{
"match": {
"nickname": "好的"
}
}
],
"should": [
{
"match": {
"sex": 1
}
}
],
"must_not":[
{
"term": {
"birthday": "1993-01-24"
}
}
]
}
},
"_source": [
"id",
"sex",
"nickname",
"desc",
"birthday"
]
}
为指定词语加权
{
"query": {
"bool": {
"should": [
{
"match": {
"desc": {
"query": "好的",
"boost": 2
}
}
},
{
"match": {
"desc": {
"query": "男孩",
"boost": 20
}
}
}
]
}
},
"_source": [
"id",
"sex",
"nickname",
"desc",
"birthday"
]
}
5 相关信息
- 博文不易，辛苦各位猿友点个关注和赞，感谢
作者:小P聊技术
原文链接:https://www.jianshu.com/p/5acd10d23d6b