Java連接Elasticsearch實戰指南

Elasticsearch是一個分散式、 RESTful風格、基於Lucene的全文搜索和分析引擎。它可以用於各種用例，例如網站搜索、日誌分析、安全情報等。

由於其性能出色和易於使用的API，越來越多的企業選擇使用Elasticsearch作為其搜索引擎。

一、連接Elasticsearch

連接Elasticsearch可以使用Java的TransportClient或Java High-Level REST Client兩種方式。

1、TransportClient

TransportClient是Elasticsearch提供的Java API，使用它可以直接連接到Elasticsearch服務。下面是一個使用TransportClient連接Elasticsearch的示例：

TransportClient client = new PreBuiltTransportClient(Settings.EMPTY)
        .addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("localhost"), 9300));

創建TransportClient實例時需要指定連接的地址和埠號。這裡我們可以使用默認的Settings和本地的9300埠。

2、Java High-Level REST Client

Java High-Level REST Client是Elasticsearch推薦的Java客戶端，它使用HTTP協議與Elasticsearch進行通信。下面是一個使用Java High-Level REST Client連接Elasticsearch的示例：

RestHighLevelClient client = new RestHighLevelClient(RestClient.builder(new HttpHost("localhost", 9200, "http")));

通過HTTP Host和埠號創建RestHighLevelClient實例。這裡我們使用本地的9200埠連接http。

二、索引操作

1、創建索引

創建索引需要指定索引名稱和索引映射。下面是一個創建名為「my_index」的索引的示例：

CreateIndexRequest request = new CreateIndexRequest("my_index");
request.mapping("doc", "field1", "type=string", "field2", "type=integer");
CreateIndexResponse response = client.indices().create(request);

創建CreateIndexRequest對象並指定索引名稱「my_index」，然後通過mapping()方法設定屬性。最後通過indices()方法的create()方法創建索引。

2、添加文檔

添加文檔需要指定索引名稱、類型和文檔ID。文檔格式一般為JSON，可以通過Map或實體類轉換得到。下面是一個添加文檔的示例：

IndexRequest request = new IndexRequest("my_index", "doc", "1");
request.source(XContentType.JSON, "field1", "value1", "field2", 2);
IndexResponse response = client.index(request);

添加文檔需要指定索引名稱、類型和文檔ID，然後通過source()方法指定文檔內容。通過index()方法添加文檔。

3、查詢文檔

查詢文檔需要指定查詢條件和索引名稱，結果一般為SearchResponse對象。下面是一個查詢文檔的示例：

SearchRequest searchRequest = new SearchRequest("my_index");
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.query(QueryBuilders.termQuery("field1", "value1"));
searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = client.search(searchRequest);

查詢文檔需要指定查詢條件和索引名稱，然後通過source()方法設置查詢條件。通過search()方法查詢文檔。

三、聚合操作

1、統計文檔數量

統計文檔數量可以使用count()方法。下面是一個統計所有文檔數量的示例：

CountRequest countRequest = new CountRequest("my_index");
CountResponse countResponse = client.count(countRequest);
long count = countResponse.getCount();

通過CountRequest對象指定索引名稱，然後通過count()方法統計文檔數量。

2、按條件聚合

聚合操作可以根據條件聚合文檔。下面是一個按「field1」欄位聚合所有文檔的示例：

SearchRequest searchRequest = new SearchRequest("my_index");
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.aggregation(AggregationBuilders.terms("by_field1").field("field1"));
searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = client.search(searchRequest);
Terms terms = searchResponse.getAggregations().get("by_field1");
for (Terms.Bucket entry : terms.getBuckets()) {
    String key = (String) entry.getKey();
    long docCount = entry.getDocCount();
}

通過AggregationBuilders對象的terms()方法指定聚合條件，結果保存在SearchResponse對象中。通過getAggregations()方法獲取所有聚合結果，然後通過get()方法獲取指定聚合結果，並使用for循環獲取聚合結果中的所有數據。

四、數據分析

1、Term Vectors

Term Vectors可以分析文檔中每個單詞的詞頻和位置等信息。下面是一個獲取Term Vectors的示例：

TermVectorsRequest request = new TermVectorsRequest("my_index", "doc", "1");
request.setFields("field1");
TermVectorsResponse response = client.termvectors(request);

通過TermVectorsRequest對象指定索引名稱、類型和文檔ID，然後通過Fields()方法設置需要分析的欄位類型。使用termvectors()方法分析文檔。

2、多關鍵字查詢

可以使用MultiMatchQuery查詢多個關鍵字。下面是一個查詢同時匹配「field1」和「field2」的文檔的示例：

SearchRequest searchRequest = new SearchRequest("my_index");
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.query(QueryBuilders.multiMatchQuery("value1", "field1", "field2"));
searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = client.search(searchRequest);

通過QueryBuilders對象的multiMatchQuery()方法指定多個查詢欄位，然後使用source()方法設置查詢條件。使用search()方法查詢文檔。

3、詞條查詢

詞條查詢可以查詢特定詞條的文檔。下面是一個查詢「field1」欄位包含「value1」的文檔的示例：

SearchRequest searchRequest = new SearchRequest("my_index");
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.query(QueryBuilders.termQuery("field1", "value1"));
searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = client.search(searchRequest);

通過QueryBuilders對象的termQuery()方法指定查詢欄位和關鍵字，然後使用source()方法設置查詢條件。使用search()方法查詢文檔。

以上是Java連接Elasticsearch的一些常用操作，本文介紹的示例代碼可以參考下列完整代碼：

完整代碼：

TransportClient client = new PreBuiltTransportClient(Settings.EMPTY)
        .addTransportAddress(new InetSocketTransportAddress(
                InetAddress.getByName("localhost"), 9300));

//創建索引
CreateIndexRequest request = new CreateIndexRequest("my_index");
request.mapping("doc", "field1", "type=string", "field2", "type=integer");
CreateIndexResponse response = client.indices().create(request);

//添加文檔
IndexRequest request = new IndexRequest("my_index", "doc", "1");
request.source(XContentType.JSON, "field1", "value1", "field2", 2);
IndexResponse response = client.index(request);

//查詢文檔
SearchRequest searchRequest = new SearchRequest("my_index");
SearchSourceBuilder searchSourceBuilder = new SearchSourceBuilder();
searchSourceBuilder.query(QueryBuilders.termQuery("field1", "value1"));
searchRequest.source(searchSourceBuilder);
SearchResponse searchResponse = client.search(searchRequest);

//統計文檔數量
CountRequest countRequest = new CountRequest("my_index");
CountResponse countResponse = client.count(countRequest);
long count = countResponse.getCount();

//聚合操作
searchSourceBuilder.aggregation(AggregationBuilders.terms("by_field1").field("field1"));
searchRequest.source(searchSourceBuilder);
searchResponse = client.search(searchRequest);
Terms terms = searchResponse.getAggregations().get("by_field1");
for (Terms.Bucket entry : terms.getBuckets()) {
    String key = (String) entry.getKey();
    long docCount = entry.getDocCount();
}

//Term Vectors
TermVectorsRequest request = new TermVectorsRequest("my_index", "doc", "1");
request.setFields("field1");
TermVectorsResponse response = client.termvectors(request);

//多關鍵字查詢
searchSourceBuilder.query(QueryBuilders.multiMatchQuery("value1", "field1", "field2"));
searchRequest.source(searchSourceBuilder);
searchResponse = client.search(searchRequest);

//詞條查詢
searchSourceBuilder.query(QueryBuilders.termQuery("field1", "value1"));
searchRequest.source(searchSourceBuilder);
searchResponse = client.search(searchRequest);