ElasticSearch笔记整理（三）：Java API使用与ES中文分词

pom.xml

使用maven工程构建ES Java API的测试项目，其用到的依赖如下：

<dependency>
    <groupId>org.elasticsearch</groupId>
    <artifactId>elasticsearch</artifactId>
    <version>2.3.0</version>
</dependency>
<dependency>
    <groupId>com.fasterxml.jackson.core</groupId>
    <artifactId>jackson-databind</artifactId>
    <version>2.7.0</version>
</dependency>
<dependency>
    <groupId>org.dom4j</groupId>
    <artifactId>dom4j</artifactId>
    <version>2.0.0</version>
</dependency>
<!--使用lombok，对于Java Bean对象，就不用手动添加getter和setter方法，在编译时，它会帮我们自动添加-->
<dependency>
    <groupId>org.projectlombok</groupId>
    <artifactId>lombok</artifactId>
    <version>1.16.10</version>
</dependency>

ES API之基本增删改查

使用junit进行测试，其使用的全局变量与setUp函数如下：

private TransportClient client;
private String index = "bigdata";   // 要操作的索引库为"bigdata"
private String type = "product";    // 要操作的类型为"product"

@Before
public void setup() throws UnknownHostException {
    // 连接的是ES集群，所以需要添加集群名称，否则无法创建客户端
    Settings settings = Settings.builder().put("cluster.name", "bigdata-08-28").build();
    client = TransportClient.builder().settings(settings).build();
    TransportAddress ta1 = new InetSocketTransportAddress(InetAddress.getByName("uplooking01"), 9300);
    TransportAddress ta2 = new InetSocketTransportAddress(InetAddress.getByName("uplooking02"), 9300);
    TransportAddress ta3 = new InetSocketTransportAddress(InetAddress.getByName("uplooking03"), 9300);
    client.addTransportAddresses(ta1, ta2, ta3);
    /*settings = client.settings();
        Map<String, String> asMap = settings.getAsMap();
        for(Map.Entry<String, String> setting : asMap.entrySet()) {
            System.out.println(setting.getKey() + "::" + setting.getValue());
        }*/
}

索引添加：JSON方式

/**
     * 注意：往es中添加数据有4种方式
     * 1.JSON
     * 2.Map
     * 3.Java Bean
     * 4.XContentBuilder
     *
     * 1.JSON方式
     */
@Test
public void testAddJSON() {
    String source = "{\"name\":\"sqoop\", \"author\": \"apache\", \"version\": \"1.4.6\"}";
    IndexResponse response = client.prepareIndex(index, type, "4").setSource(source).get();
    System.out.println(response.isCreated());
}

索引添加：Map方式

/**
     * 添加数据：
     * 2.Map方式
     */
@Test
public void testAddMap() {
    Map<String, Object> source = new HashMap<String, Object>();
    source.put("name", "flume");
    source.put("author", "Cloudera");
    source.put("version", "1.8.0");
    IndexResponse response = client.prepareIndex(index, type, "5").setSource(source).get();
    System.out.println(response.isCreated());
}

索引添加：Java Bean方式

/**
     * 添加数据：
     * 3.Java Bean方式
     *
     * 如果不将对象转换为json字符串，则会报下面的异常：
     * The number of object passed must be even but was [1]
     */
@Test
public void testAddObj() throws JsonProcessingException {
    Product product = new Product("kafka", "linkedIn", "0.10.0.1", "kafka.apache.org");
    ObjectMapper objectMapper = new ObjectMapper();
    String json = objectMapper.writeValueAsString(product);
    System.out.println(json);
    IndexResponse response = client.prepareIndex(index, type, "6").setSource(json).get();
    System.out.println(response.isCreated());
}

索引添加：XContentBuilder方式

/**
     * 添加数据：
     * 4.XContentBuilder方式
     */
@Test
public void testAddXContentBuilder() throws IOException {
    XContentBuilder source = XContentFactory.jsonBuilder();
    source.startObject()
        .field("name", "redis")
        .field("author", "redis")
        .field("version", "3.2.0")
        .field("url", "redis.cn")
        .endObject();
    IndexResponse response = client.prepareIndex(index, type, "7").setSource(source).get();
    System.out.println(response.isCreated());
}

索引查询

/**
     * 查询具体的索引信息
     */
@Test
public void testGet() {
    GetResponse response = client.prepareGet(index, type, "6").get();
    Map<String, Object> map = response.getSource();
    /*for(Map.Entry<String, Object> me : map.entrySet()) {
            System.out.println(me.getKey() + "=" + me.getValue());
        }*/
    // lambda表达式，jdk 1.8之后
    map.forEach((k, v) -> System.out.println(k + "=" + v));
    //        map.keySet().forEach(key -> System.out.println(key + "xxx"));
}

索引更新

/**
     * 局部更新操作与curl的操作是一致的
     * curl -XPOST http://uplooking01:9200/bigdata/product/AWA184kojrSrzszxL-Zs/_update -d‘ {"doc":{"name":"sqoop", "author":"apache"}}‘
     *
     * 做全局更新的时候，也不用prepareUpdate，而直接使用prepareIndex
     */
@Test
public void testUpdate() throws Exception {
    /*String source = "{\"doc\":{\"url\": \"http://flume.apache.org\"}}";
        UpdateResponse response = client.prepareUpdate(index, type, "4").setSource(source.getBytes()).get();*/
    // 使用下面这种方式也是可以的
    String source = "{\"url\": \"http://flume.apache.org\"}";
    UpdateResponse response = client.prepareUpdate(index, type, "4").setDoc(source.getBytes()).get();
    System.out.println(response.getVersion());
}

索引删除

/**
     * 删除操作
     */
@Test
public void testDelete() {
    DeleteResponse response = client.prepareDelete(index, type, "5").get();
    System.out.println(response.getVersion());
}

批量操作

/**
     * 批量操作
     */
@Test
public void testBulk() {
    IndexRequestBuilder indexRequestBuilder = client.prepareIndex(index, type, "8")
        .setSource("{\"name\":\"elasticsearch\", \"url\":\"http://www.elastic.co\"}");
    UpdateRequestBuilder updateRequestBuilder = client.prepareUpdate(index, type, "1").setDoc("{\"url\":\"http://hadoop.apache.org\"}");
    BulkRequestBuilder bulk = client.prepareBulk();
    BulkResponse bulkResponse = bulk.add(indexRequestBuilder).add(updateRequestBuilder).get();
    Iterator<BulkItemResponse> it = bulkResponse.iterator();
    while(it.hasNext()) {
        BulkItemResponse response = it.next();
        System.out.println(response.getId() + "<--->" + response.getVersion());
    }
}

获取索引记录数

/**
     * 获取索引记录数
     */
@Test
public void testCount() {
    CountResponse response = client.prepareCount(index).get();
    System.out.println("索引记录数：" + response.getCount());
}

ES API之高级查询

基于junit进行测试，其用到的setUp函数和showResult函数如下：

全局变量与setUp：

private TransportClient client;
private String index = "bigdata";
private String type = "product";
private String[] indics = {"bigdata", "bank"};

@Before
public void setUp() throws UnknownHostException {
    Settings settings = Settings.builder().put("cluster.name", "bigdata-08-28").build();
    client = TransportClient.builder().settings(settings).build();
    TransportAddress ta1 = new InetSocketTransportAddress(InetAddress.getByName("uplooking01"), 9300);
    TransportAddress ta2 = new InetSocketTransportAddress(InetAddress.getByName("uplooking02"), 9300);
    TransportAddress ta3 = new InetSocketTransportAddress(InetAddress.getByName("uplooking03"), 9300);
    client.addTransportAddresses(ta1, ta2, ta3);
}

showResult：

/**
     * 格式化输出查询结果
     * @param response
     */
private void showResult(SearchResponse response) {
    SearchHits searchHits = response.getHits();
    float maxScore = searchHits.getMaxScore();  // 查询结果中的最大文档得分
    System.out.println("maxScore: " + maxScore);
    long totalHits = searchHits.getTotalHits(); // 查询结果记录条数
    System.out.println("totalHits: " + totalHits);
    SearchHit[] hits = searchHits.getHits();    // 查询结果
    System.out.println("当前返回结果记录条数：" + hits.length);
    for (SearchHit hit : hits) {
        long version = hit.version();
        String id = hit.getId();
        String index = hit.getIndex();
        String type = hit.getType();
        float score = hit.getScore();
        System.out.println("===================================================");
        String source = hit.getSourceAsString();
        System.out.println("version: " + version);
        System.out.println("id: " + id);
        System.out.println("index: " + index);
        System.out.println("type: " + type);
        System.out.println("score: " + score);
        System.out.println("source: " + source);
    }
}

ES查询类型说明

查询类型有如下4种：

query and fetch(速度最快)(返回N倍数据量)
query then fetch（默认的搜索方式）
DFS query and fetch
DFS query then fetch(可以更精确控制搜索打分和排名。)

查看API的注释如下：

/**
     * Same as {@link #QUERY_THEN_FETCH}, except for an initial scatter phase which goes and computes the distributed
     * term frequencies for more accurate scoring.
     */
DFS_QUERY_THEN_FETCH((byte) 0),
/**
     * The query is executed against all shards, but only enough information is returned (not the document content).
     * The results are then sorted and ranked, and based on it, only the relevant shards are asked for the actual
     * document content. The return number of hits is exactly as specified in size, since they are the only ones that
     * are fetched. This is very handy when the index has a lot of shards (not replicas, shard id groups).
     */
QUERY_THEN_FETCH((byte) 1),
/**
     * Same as {@link #QUERY_AND_FETCH}, except for an initial scatter phase which goes and computes the distributed
     * term frequencies for more accurate scoring.
     */
DFS_QUERY_AND_FETCH((byte) 2),
/**
     * The most naive (and possibly fastest) implementation is to simply execute the query on all relevant shards
     * and return the results. Each shard returns size results. Since each shard already returns size hits, this
     * type actually returns size times number of shards results back to the caller.
     */
QUERY_AND_FETCH((byte) 3),

关于DFS的说明：

DFS是什么缩写？
这个D可能是Distributed，F可能是frequency的缩写，至于S可能是Scatter的缩写，整个单词可能是分布式词频率和
文档频率散发的缩写。

初始化散发是一个什么样的过程？
从es的官方网站我们可以发现，初始化散发其实就是在进行真正的查询之前，先把各个分片的词频率和文档频率收集一
下，然后进行词搜索的时候，各分片依据全局的词频率和文档频率进行搜索和排名。显然如果使用
DFS_QUERY_THEN_FETCH这种查询方式，效率是最低的，因为一个搜索，可能要请求3次分片。但，使用DFS方法，搜索
精度应该是最高的。

总结：

总结一下，从性能考虑QUERY_AND_FETCH是最快的，DFS_QUERY_THEN_FETCH是最慢的。从搜索的准确度来说，DFS要
比非DFS的准确度更高。

精确查询

/**
     * 1.精确查询
     * termQuery
     * term就是一个字段
     */
@Test
public void testSearch1() {
    SearchRequestBuilder searchQuery = client.prepareSearch(indics)    // 在prepareSearch()的参数为索引库列表，意为要从哪些索引库中进行查询
        .setSearchType(SearchType.DEFAULT)  // 设置查询类型，有QUERY_AND_FETCH  QUERY_THEN_FETCH  DFS_QUERY_AND_FETCH  DFS_QUERY_THEN_FETCH
        .setQuery(QueryBuilders.termQuery("author", "apache"))// 设置相应的query，用于检索，termQuery的参数说明：name是doc中的具体的field，value就是要找的具体的值
        ;
    // 如果上面不加查询条件，则会查询所有
    SearchResponse response = searchQuery.get();

    showResult(response);
}

模糊查询

/**
     * 2.模糊查询
     * prefixQuery
     */
@Test
public void testSearch2() {
    SearchResponse response = client.prepareSearch(indics).setSearchType(SearchType.QUERY_THEN_FETCH)
        .setQuery(QueryBuilders.prefixQuery("name", "h"))
        .get();
    showResult(response);
}

分页查询

/**
     * 3.分页查询
     * 查询索引库bank中
     * 年龄在(25, 35]之间的数据信息
     *
     * 分页算法：
     *      查询的第几页，每一页显示几条
     *          每页显示10条记录
     *
     *      查询第4页的内容
     *          setFrom(30=(4-1)*size)
     *          setSize(10)
     *       所以第N页的起始位置：(N - 1) * pageSize
     */
@Test
public void testSearch3() {
    // 注意QUERY_THEN_FETCH和注意QUERY_AND_FETCH返回的记录数不一样，前者默认10条，后者是50条（5个分片）
    SearchResponse response = client.prepareSearch(indics).setSearchType(SearchType.DFS_QUERY_THEN_FETCH)
        .setQuery(QueryBuilders.rangeQuery("age").gt(25).lte(35))
        // 下面setFrom和setSize用于设置查询结果进行分页
        .setFrom(0)
        .setSize(5)
        .get();
    showResult(response);
}

高亮显示查询

/**
     * 4.高亮显示查询
     * 获取数据，
     *  查询apache，不仅在author拥有，也可以在url，在name中也可能拥有
     *  author or url   --->booleanQuery中的should操作
     *      如果是and的类型--->booleanQuery中的must操作
     *      如果是not的类型--->booleanQuery中的mustNot操作
     *  使用的match操作，其实就是使用要查询的keyword和对应字段进行完整匹配，是否相等，相等返回
     */
@Test
public void testSearch4() {
    SearchResponse response = client.prepareSearch(indics).setSearchType(SearchType.DEFAULT)
        //                .setQuery(QueryBuilders.multiMatchQuery("apache", "author", "url"))
        //                .setQuery(QueryBuilders.regexpQuery("url", ".*apache.*"))
        //                .setQuery(QueryBuilders.termQuery("author", "apache"))
        .setQuery(QueryBuilders.boolQuery()
                  .should(QueryBuilders.regexpQuery("url", ".*apache.*"))
                  .should(QueryBuilders.termQuery("author", "apache")))
        // 设置高亮显示--->设置相应的前置标签和后置标签
        .setHighlighterPreTags("<span color=‘blue‘ size=‘18px‘>")
        .setHighlighterPostTags("</span>")
        // 哪个字段要求高亮显示
        .addHighlightedField("author")
        .addHighlightedField("url")
        .get();
    SearchHits searchHits = response.getHits();
    float maxScore = searchHits.getMaxScore();  // 查询结果中的最大文档得分
    System.out.println("maxScore: " + maxScore);
    long totalHits = searchHits.getTotalHits(); // 查询结果记录条数
    System.out.println("totalHits: " + totalHits);
    SearchHit[] hits = searchHits.getHits();    // 查询结果
    System.out.println("当前返回结果记录条数：" + hits.length);
    for(SearchHit hit : hits) {
        System.out.println("========================================================");
        Map<String, HighlightField> highlightFields = hit.getHighlightFields();
        for(Map.Entry<String , HighlightField> me : highlightFields.entrySet()) {
            System.out.println("--------------------------------------");
            String key = me.getKey();
            HighlightField highlightField = me.getValue();
            String name = highlightField.getName();
            System.out.println("key: " + key + ", name: " + name);
            Text[] texts = highlightField.fragments();
            String value = "";
            for(Text text : texts) {
                // System.out.println("text: " + text.toString());
                value += text.toString();
            }
            System.out.println("value: " + value);
        }
    }
}

排序查询

/**
     * 5.排序查询
     * 对结果集进行排序
     *  balance（收入）由高到低
     */
@Test
public void testSearch5() {
    // 注意QUERY_THEN_FETCH和注意QUERY_AND_FETCH返回的记录数不一样，前者默认10条，后者是50条（5个分片）
    SearchResponse response = client.prepareSearch(indics).setSearchType(SearchType.DFS_QUERY_THEN_FETCH)
        .setQuery(QueryBuilders.rangeQuery("age").gt(25).lte(35))
        .addSort("balance", SortOrder.DESC)
        // 下面setFrom和setSize用于设置查询结果进行分页
        .setFrom(0)
        .setSize(5)
        .get();
    showResult(response);
}

聚合查询：计算平均值

/**
     * 6.聚合查询：计算平均值
     */
@Test
public void testSearch6() {
    indics = new String[]{"bank"};
    // 注意QUERY_THEN_FETCH和注意QUERY_AND_FETCH返回的记录数不一样，前者默认10条，后者是50条（5个分片）
    SearchResponse response = client.prepareSearch(indics).setSearchType(SearchType.DFS_QUERY_THEN_FETCH)
        .setQuery(QueryBuilders.rangeQuery("age").gt(25).lte(35))
        /*
                    select avg(age) as avg_name from person;
                    那么这里的avg("balance")--->就是返回结果avg_name这个别名
                 */
        .addAggregation(AggregationBuilders.avg("avg_balance").field("balance"))
        .addAggregation(AggregationBuilders.max("max").field("balance"))
        .get();
    //        System.out.println(response);
    /*
            response中包含的Aggregations
                "aggregations" : {
                    "max" : {
                      "value" : 49741.0
                    },
                    "avg_balance" : {
                      "value" : 25142.137373737372
                    }
                  }
                  则一个aggregation为：
                  {
                      "value" : 49741.0
                    }
         */
    Aggregations aggregations = response.getAggregations();
    List<Aggregation> aggregationList = aggregations.asList();
    for(Aggregation aggregation : aggregationList) {
        System.out.println("========================================");
        String name = aggregation.getName();
        // Map<String, Object> map = aggregation.getMetaData();
        System.out.println("name: " + name);
        // System.out.println(map);
        Object obj = aggregation.getProperty("value");
        System.out.println(obj);
    }
    /*Aggregation avgBalance = aggregations.get("avg_balance");
        Object obj = avgBalance.getProperty("value");
        System.out.println(obj);*/
}

ES中文分词之集成IK分词

如果我们的数据包含中文，而在查询时希望可以支持对中文进行分词搜索，那么ES本身依赖于Lucene的分词对中文就不佳了，这时就可以考虑使用其它分词方法，如这里要说明的IK中文分词，其集成到ES的步骤如下：

  1)下载地址：
    https://github.com/medcl/elasticsearch-analysis-ik
  2)使用maven对源代码进行编译(mvn clean install -DskipTests)（package）
  3)把编译后的target/releases下的zip文件拷贝到   ES_HOME/plugins/analysis-ik目录下面，然后解压
  4)把下载的ik插件中的conf/ik目录拷贝到ES_HOME/config下
  5)修改ES_HOME/config/elasticsearch.yml文件，添加index.analysis.analyzer.default.type: ik
  (把IK设置为默认分词器,这一步是可选的)
  6)重启es服务
  7)测试分词效果

需要说明的是，数据需要重新插入，并使用ik分词，即需要重新构建期望使用中文分词IK的索引库。

测试代码如下：

package cn.xpleaf.bigdata.elasticsearch;

import org.elasticsearch.action.search.SearchRequestBuilder;
import org.elasticsearch.action.search.SearchResponse;
import org.elasticsearch.action.search.SearchType;
import org.elasticsearch.client.transport.TransportClient;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.text.Text;
import org.elasticsearch.common.transport.InetSocketTransportAddress;
import org.elasticsearch.common.transport.TransportAddress;
import org.elasticsearch.index.query.QueryBuilders;
import org.elasticsearch.search.SearchHit;
import org.elasticsearch.search.SearchHits;
import org.elasticsearch.search.aggregations.Aggregation;
import org.elasticsearch.search.aggregations.AggregationBuilders;
import org.elasticsearch.search.aggregations.Aggregations;
import org.elasticsearch.search.highlight.HighlightField;
import org.elasticsearch.search.sort.SortOrder;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;

import java.net.InetAddress;
import java.net.UnknownHostException;
import java.util.List;
import java.util.Map;

/**
 * 使用Java API来操作es集群
 * Transport
 * 代表了一个集群
 * 我们客户端和集群通信是使用TransportClient
 * <p>
 * 使用prepareSearch来完成全文检索之
 *  中文分词
 */
public class ElasticSearchTest3 {

    private TransportClient client;
    private String index = "bigdata";
    private String type = "product";
    private String[] indics = {"chinese"};

    @Before
    public void setUp() throws UnknownHostException {
        Settings settings = Settings.builder().put("cluster.name", "bigdata-08-28").build();
        client = TransportClient.builder().settings(settings).build();
        TransportAddress ta1 = new InetSocketTransportAddress(InetAddress.getByName("uplooking01"), 9300);
        TransportAddress ta2 = new InetSocketTransportAddress(InetAddress.getByName("uplooking02"), 9300);
        TransportAddress ta3 = new InetSocketTransportAddress(InetAddress.getByName("uplooking03"), 9300);
        client.addTransportAddresses(ta1, ta2, ta3);
    }

    /**
     * 中文分词的操作
     * 1.查询以"中"开头的数据，有两条
     * 2.查询以“中国”开头的数据，有0条
     * 3.查询包含“烂”的数据，有1条
     * 4.查询包含“烂摊子”的数据，有0条
     * 分词：
     *      为什么我们搜索China is the greatest country~
     *                 中文：中国最牛逼
     *
     *                 ×××
     *                      中华
     *                      人民
     *                      共和国
     *                      中华人民
     *                      人民共和国
     *                      华人
     *                      共和
     *      特殊的中文分词法：
     *          庖丁解牛
     *          IK分词法
     *          搜狗分词法
     */
    @Test
    public void testSearch1() {
        SearchResponse response = client.prepareSearch(indics)    // 在prepareSearch()的参数为索引库列表，意为要从哪些索引库中进行查询
                .setSearchType(SearchType.DEFAULT)  // 设置查询类型，有QUERY_AND_FETCH  QUERY_THEN_FETCH  DFS_QUERY_AND_FETCH  DFS_QUERY_THEN_FETCH
                //.setQuery(QueryBuilders.prefixQuery("content", "烂摊子"))// 设置相应的query，用于检索，termQuery的参数说明：name是doc中的具体的field，value就是要找的具体的值
//                .setQuery(QueryBuilders.regexpQuery("content", ".*烂摊子.*"))
                .setQuery(QueryBuilders.prefixQuery("content", "中国"))
                .get();
        showResult(response);
    }

    /**
     * 格式化输出查询结果
     * @param response
     */
    private void showResult(SearchResponse response) {
        SearchHits searchHits = response.getHits();
        float maxScore = searchHits.getMaxScore();  // 查询结果中的最大文档得分
        System.out.println("maxScore: " + maxScore);
        long totalHits = searchHits.getTotalHits(); // 查询结果记录条数
        System.out.println("totalHits: " + totalHits);
        SearchHit[] hits = searchHits.getHits();    // 查询结果
        System.out.println("当前返回结果记录条数：" + hits.length);
        for (SearchHit hit : hits) {
            long version = hit.version();
            String id = hit.getId();
            String index = hit.getIndex();
            String type = hit.getType();
            float score = hit.getScore();
            System.out.println("===================================================");
            String source = hit.getSourceAsString();
            System.out.println("version: " + version);
            System.out.println("id: " + id);
            System.out.println("index: " + index);
            System.out.println("type: " + type);
            System.out.println("score: " + score);
            System.out.println("source: " + source);
        }
    }

    @After
    public void cleanUp() {
        client.close();
    }
}

相关测试代码已上传到GitHub：https://github.com/xpleaf/elasticsearch-study

原文：http://blog.51cto.com/xpleaf/2096854