使用以下脚本我设置elasticsearch来索引文章。每篇文章都有一份作者列表：
#!/usr/bin/env bash
HOST='http://localhost:9200/articles'
curl -XPUT ${HOST}
curl -XPUT "${HOST}/article/_mapping" -d '{
"article" : {
"properties" : {
"authors" : {"type":"string", "index_name":"author", "index": "not_analyzed"}
}
}
}'
curl -XPOST "${HOST}/article/1" -d '{
"authors" : ["Albert","Wolfgang","Richard","Murray"],
"message" : "Blabla" }'
curl -XPOST "${HOST}/article/2" -d '{
"authors" : ["Albert","Richard"],
"message" : "Blublu" }'
curl -XPOST "${HOST}/article/3" -d '{
"authors" : ["Albert"],
"message" : "Bleble" }'
我想要做的是过滤掉所有不在给定列表中的作者的文章。我尝试了以下查询：
curl -XGET "${HOST}/_search?pretty=true" -d '{
"query": {
"constant_score": {
"filter": {
"terms": {
"authors": ["Albert","Richard","Erwin"],
"execution": "or"
}
}
}
}
}'
然而，这将返回所有三篇文章作为点击。但我确实希望过滤掉第1条，因为它有一些作者["Wolfgang", "Murray"]，这些作者不在给定作者["Albert", "Richard","Erwin"]的列表中。这可以用弹性搜索来实现吗？
答案 0 :(得分：1)
棘手的一个。如果我理解得很清楚，您需要查找作者字段中仅包含 这三个值中的一个（或更多）的文档
根据elasticsearch权威指南here，检查字段是否包含 某些值非常昂贵。
有一个丑陋的解决方案受到equality paragraph的启发，它会使用bool过滤器的组合和添加author_count字段。
您必须检查是否：
这对我来说不可行。
答案 1 :(得分：1)
正如Tom83所建议的那样，在弹性搜索权威指南链接中，这可以通过在索引时添加authors_counts属性来完成。像这样：
curl -XPOST "${HOST}/article/1" -d '{
"authors": [
"Albert",
"Wolfgang",
"Richard",
"Murray"
],
"authors_counts": 4,
"message": "Blabla"
}'
curl -XPOST "${HOST}/article/2" -d '{
"authors": [
"Albert",
"Richard"
],
"authors_counts": 2,
"message": "Blublu"
}'
curl -XPOST "${HOST}/article/3" -d '{
"authors": [
"Albert"
],
"authors_counts": 1,
"message": "Bleble"
}'
然后，您可以按照指南中的建议创建查询。它非常冗长，所以我决定生成它。
(ns query-gen
(:require [clojure.data.json :as json]
[clojure.math.combinatorics :refer [subsets]]))
(defn gen-filter [items]
(let [terms (map (fn [term] { "term" { "authors" term } }) items)
terms_count { "term" { "authors_counts" (count items) }}]
{ "bool" { "must" (cons terms_count terms)}}))
(defn gen-query [names]
(let [subsets (rest (subsets names))
filters (map gen-filter subsets)]
{"query" { "filtered" { "filter" { "or" filters }}}}))
(defn -main [& args]
(let [ query (gen-query ["Albert" "Richard" "Erwin"])
json (json/write-str query)]
(println json)))
这会生成如下所示的查询：
{
"query": {
"filtered": {
"filter": {
"or": [
{
"bool": {
"must": [
{
"term": {
"authors_counts": 1
}
},
{
"term": {
"authors": "Albert"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 1
}
},
{
"term": {
"authors": "Richard"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 1
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 2
}
},
{
"term": {
"authors": "Albert"
}
},
{
"term": {
"authors": "Richard"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 2
}
},
{
"term": {
"authors": "Albert"
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 2
}
},
{
"term": {
"authors": "Richard"
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 3
}
},
{
"term": {
"authors": "Albert"
}
},
{
"term": {
"authors": "Richard"
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
}
]
}
}
}
}
如果使用如下：
curl -XGET "${HOST}/_search?pretty=true" -d '{
"query": {
"filtered": {
"filter": {
"or": [
{
"bool": {
"must": [
{
"term": {
"authors_counts": 1
}
},
{
"term": {
"authors": "Albert"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 1
}
},
{
"term": {
"authors": "Richard"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 1
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 2
}
},
{
"term": {
"authors": "Albert"
}
},
{
"term": {
"authors": "Richard"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 2
}
},
{
"term": {
"authors": "Albert"
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 2
}
},
{
"term": {
"authors": "Richard"
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"authors_counts": 3
}
},
{
"term": {
"authors": "Albert"
}
},
{
"term": {
"authors": "Richard"
}
},
{
"term": {
"authors": "Erwin"
}
}
]
}
}
]
}
}
}
}'
它会返回我认为您期望的结果：
{
"took": 5,
"timed_out": false,
"_shards": {
"total": 5,
"successful": 5,
"failed": 0
},
"hits": {
"total": 2,
"max_score": 1.0,
"hits": [
{
"_index": "articles",
"_type": "article",
"_id": "3",
"_score": 1.0,
"_source": {
"authors": [
"Albert"
],
"authors_counts": 1,
"message": "Bleble"
}
},
{
"_index": "articles",
"_type": "article",
"_id": "2",
"_score": 1.0,
"_source": {
"authors": [
"Albert",
"Richard"
],
"authors_counts": 2,
"message": "Blublu"
}
}
]
}
}
我不确定这是不是一个好主意，但非常有趣。希望它有所帮助！