ES

理解索引結構

邏輯結構部分是一個倒排索引表：
一、將要搜索的文檔內容分詞，全部不重複的詞組成分詞列表。
二、將搜索的文檔最終以Document方式存儲起來。
三、每一個詞和docment都有關聯。

下面是Elasticsearch與MySQL數據庫邏輯結構概念的對比

索引(index) =數據庫(databases)
類型(type) =表(table)
文檔(document)= 行(row)

字符串類型

text
當一個字段是要被全文搜索的，好比Email內容、產品描述，應該使用text類型。設置text類型之後，字段內容會被分析，在生成倒排索引之前，字符串會被分析器分紅一個一個詞項。text類型的字段不用於排序，不多用於聚合。
keyword
keyword類型適用於索引結構化的字段，好比email地址、主機名、狀態碼和標籤。若是字段須要進行過濾(好比查找已發佈博客中status屬性爲published的文章)、排序、聚合。keyword類型的字段只能經過精確值搜索到。

JavaRest 高級客戶端簡介

elasticsearch 存在三種Java客戶端。

1. Transport Client
2. Java Low Level Rest Client（低級rest客戶端）
3. Java High Level REST Client（高級rest客戶端）
   這三者的區別是：
   TransportClient沒有使用RESTful風格的接口，而是二進制的方式傳輸數據。
   ES官方推出了Java Low Level REST Client,它支持RESTful。可是缺點是由於
   TransportClient的使用者把代碼遷移到Low Level REST Client的工做量比較大。
   ES官方推出Java High Level REST Client,它是基於Java Low Level REST Client的封
   裝，而且API接收參數和返回值和TransportClient是同樣的，使得代碼遷移變得容易
   而且支持了RESTful的風格，兼容了這兩種客戶端的優勢。強烈建議ES5及其之後的
   版本使用Java High Level REST Client。
   準備工做：新建工程，引入依賴

<dependency>
<groupId>org.elasticsearch.client</groupId>
<artifactId>elasticsearch‐rest‐high‐level‐client</artifactId>
<version>6.5.3</version>
</dependency>

新增和修改數據

插入單條數據：

HttpHost : url地址封裝
RestClientBuilder： rest客戶端構建器

IndexRequest： 新增或修改請求
IndexResponse：新增或修改的響應結果

//1.鏈接rest接口
HttpHost http=new HttpHost("127.0.0.1",9200,"http");
RestClientBuilder builder= RestClient.builder(http);//rest構建器
RestHighLevelClient restHighLevelClient=new RestHighLevelClient(builder);//高級客戶端對象
//2.封裝請求對象
IndexRequest indexRequest=new IndexRequest("sku","doc","3");
Map skuMap =new HashMap();
skuMap.put("name","華爲p30pro");
skuMap.put("saleNum",101021);
skuMap.put("commentNum",10102321);
Map spec=new HashMap();
spec.put("網絡制式","移動4G");
spec.put("屏幕尺寸","5");
skuMap.put("spec",spec);
indexRequest.source(skuMap);
//3.獲取響應結果
IndexResponse response = restHighLevelClient.index(indexRequest,RequestOptions.DEFAULT);
int status = response.status().getStatus();
System.out.println(status);
restHighLevelClient.close();

若是ID不存在則新增，若是ID存在則修改。
批處理請求：
BulkRequest： 批量請求（用於增刪改操做）
BulkResponse：批量請求（用於增刪改操做）

//1.鏈接rest接口
HttpHost http=new HttpHost("127.0.0.1",9200,"http");
RestClientBuilder builder= RestClient.builder(http);//rest構建器
RestHighLevelClient restHighLevelClient=new RestHighLevelClient(builder);//高級客戶端對象
//2.封裝請求對象
BulkRequest bulkRequest=new BulkRequest();
IndexRequest indexRequest=new IndexRequest("sku","doc","4");
Map skuMap =new HashMap();
skuMap.put("name","華爲p30pro 火爆上市");
skuMap.put("brandName","華爲");
Map spec=new HashMap();
spec.put("網絡制式","移動4G");
spec.put("屏幕尺寸","5");
skuMap.put("spec",spec);
indexRequest.source(skuMap);
bulkRequest.add(indexRequest);//能夠屢次添加
//3.獲取響應結果
BulkResponse response =restHighLevelClient.bulk(bulkRequest,RequestOptions.DEFAULT);
int status = response.status().getStatus();
System.out.println(status);
String message = response.buildFailureMessage();
System.out.println(message);
restHighLevelClient.close();

匹配查詢

SearchRequest： 查詢請求對象

SearchResponse：查詢響應對象

SearchSourceBuilder：查詢源構建器

MatchQueryBuilder：匹配查詢構建器

示例：查詢商品名稱包含手機的記錄。

//1.鏈接rest接口
HttpHost http=new HttpHost("127.0.0.1",9200,"http");
RestClientBuilder builder= RestClient.builder(http);//rest構建器
RestHighLevelClient restHighLevelClient=new RestHighLevelClient(builder);//高級客戶端對象
//2.封裝查詢請求
SearchRequest searchRequest=new SearchRequest("sku");
searchRequest.types("doc"); //設置查詢的類型
SearchSourceBuilder searchSourceBuilder=new SearchSourceBuilder();
MatchQueryBuilder matchQueryBuilder= QueryBuilders.matchQuery("name","手機");
searchSourceBuilder.query(matchQueryBuilder);
searchRequest.source(searchSourceBuilder);
//3.獲取查詢結果
SearchResponse searchResponse = restHighLevelClient.search(searchRequest,RequestOptions.DEFAULT);
SearchHits searchHits = searchResponse.getHits();
long totalHits = searchHits.getTotalHits();
System.out.println("記錄數："+totalHits);
SearchHit[] hits = searchHits.getHits();
for(SearchHit hit:hits){
    String source = hit.getSourceAsString();
    System.out.println(source);
}
restHighLevelClient.close();

布爾與詞條查詢

BoolQueryBuilder：布爾查詢構建器
TermQueryBuilder：詞條查詢構建器
QueryBuilders：查詢構建器工廠

示例：查詢名稱包含手機的，而且品牌爲小米的記錄

//1.鏈接rest接口 同上......
//2.封裝查詢請求
SearchRequest searchRequest=new SearchRequest("sku");
searchRequest.types("doc"); //設置查詢的類型
SearchSourceBuilder searchSourceBuilder=new SearchSourceBuilder();
BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();//布爾查詢構建器
MatchQueryBuilder matchQueryBuilder= QueryBuilders.matchQuery("name","手機");
boolQueryBuilder.must(matchQueryBuilder);
TermQueryBuilder termQueryBuilder = QueryBuilders.termQuery("brandName","小米");
boolQueryBuilder.must(termQueryBuilder);
searchSourceBuilder.query(boolQueryBuilder);
searchRequest.source(searchSourceBuilder);
//3.獲取查詢結果 同上......

過濾查詢

過濾是針對搜索的結果進行過濾，過濾器主要判斷的是文檔是否匹配，不去計算和判斷文檔的匹配度得分，因此過濾器性能比查詢要高，且方便緩存，推薦儘可能使用過濾器去實現查詢或者過濾器和查詢共同使用。

示例：篩選品牌爲小米的記錄

//1.鏈接rest接口 同上.....
//2.封裝查詢請求
SearchRequest searchRequest=new SearchRequest("sku");
searchRequest.types("doc"); //設置查詢的類型
SearchSourceBuilder searchSourceBuilder=new SearchSourceBuilder();
BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();//布爾查詢構建器
TermQueryBuilder termQueryBuilder = QueryBuilders.termQuery("brandName","小米");
boolQueryBuilder.filter(termQueryBuilder);
searchSourceBuilder.query(boolQueryBuilder);
searchRequest.source(searchSourceBuilder);
//3.獲取查詢結果 同上....

分組（聚合）查詢

示例：按分組名稱聚合查詢，統計每一個分組的數量

GET /sku/_search
{
        "size" : 0,
        "aggs" : {
            "sku_category" : {
                "terms" : {
                     "field" : "categoryName"
                  }
               }
      }
}

size爲0 不會將數據查詢出來，目的是讓查詢更快。

結果：

{
    "took": 6,
    "timed_out": false,
    "_shards": {
        "total": 5,
        "successful": 5,
        "skipped": 0,
        "failed": 0
    },
    "hits": {
        "total": 5,
        "max_score": 0,
        "hits": []
    },
    "aggregations": {
        "sku_category": {
            "doc_count_error_upper_bound": 0,
            "sum_other_doc_count": 0,
            "buckets": [
                {
                "key": "手機",
                "doc_count": 3
                },
                {
                "key": "電視",    
                "doc_count": 2
                }
            ]
    }
}
}

AggregationBuilders：聚合構建器工廠
TermsAggregationBuilder：詞條聚合構建器
Aggregations：分組結果封裝
Terms.Bucket： 桶

示例：按商品分類分組查詢，求出每一個分類的文檔數

//1.鏈接rest接口
HttpHost http=new HttpHost("127.0.0.1",9200,"http");
RestClientBuilder builder= RestClient.builder(http);//rest構建器
RestHighLevelClient restHighLevelClient=new
RestHighLevelClient(builder);//高級客戶端對象
//2.封裝查詢請求
SearchRequest searchRequest=new SearchRequest("sku");
searchRequest.types("doc"); //設置查詢的類型
SearchSourceBuilder searchSourceBuilder=new SearchSourceBuilder();
TermsAggregationBuilder termsAggregationBuilder =AggregationBuilders.terms("sku_category").field("categoryName");  //這個field是真正的列，terms是定義的臨時別名
searchSourceBuilder.aggregation(termsAggregationBuilder);
searchSourceBuilder.size(0);//不會將數據查詢出來，目的是讓查詢更快。
searchRequest.source(searchSourceBuilder);
//3.獲取查詢結果
SearchResponse searchResponse = restHighLevelClient.search(searchRequest,
RequestOptions.DEFAULT);
Aggregations aggregations = searchResponse.getAggregations();
Map<String, Aggregation> asMap = aggregations.getAsMap();
Terms terms = (Terms) asMap.get("sku_category");
List<? extends Terms.Bucket> buckets = terms.getBuckets();
for(Terms.Bucket bucket:buckets){
System.out.println(bucket.getKeyAsString()+":"+ bucket.getDocCount()
);
}
restHighLevelClient.close();