elastic search java api入门

 

 

参考资料：https://www.elastic.co/guide/en/elasticsearch/client/java-api/2.4/index.html

注：es版本为2.4.0

 

1.下载安装：

下载地址：https://www.elastic.co/downloads/elasticsearch

下载之后解压，运行 bin/elasticsearch 即可

 

2.maven依赖：

<dependency>

    <groupId>org.elasticsearch</groupId>

    <artifactId>elasticsearch</artifactId>

    <version>2.4.0</version>

</dependency>

<dependency>

  <groupId>com.alibaba</groupId>

  <artifactId>fastjson</artifactId>

  <version>1.2.8</version>

</dependency>

 

3.具体的demo:

package com.tch.test.elasticsearch.test;

import java.net.InetAddress;

import java.net.UnknownHostException;

import java.util.ArrayList;

import java.util.Arrays;

import java.util.Date;

import java.util.List;

import java.util.concurrent.ExecutionException;

import java.util.concurrent.TimeUnit;

import org.apache.commons.collections4.CollectionUtils;

import org.apache.commons.lang3.StringUtils;

import org.elasticsearch.action.admin.indices.create.CreateIndexRequestBuilder;

import org.elasticsearch.action.admin.indices.delete.DeleteIndexRequestBuilder;

import org.elasticsearch.action.admin.indices.exists.indices.IndicesExistsResponse;

import org.elasticsearch.action.delete.DeleteResponse;

import org.elasticsearch.action.get.GetResponse;

import org.elasticsearch.action.get.MultiGetItemResponse;

import org.elasticsearch.action.get.MultiGetResponse;

import org.elasticsearch.action.index.IndexResponse;

import org.elasticsearch.action.search.SearchResponse;

import org.elasticsearch.action.update.UpdateRequest;

import org.elasticsearch.action.update.UpdateResponse;

import org.elasticsearch.client.transport.TransportClient;

import org.elasticsearch.common.transport.InetSocketTransportAddress;

import org.elasticsearch.common.xcontent.XContentBuilder;

import org.elasticsearch.common.xcontent.XContentFactory;

import org.elasticsearch.index.query.BoolQueryBuilder;

import org.elasticsearch.index.query.QueryBuilders;

import org.elasticsearch.search.SearchHit;

import org.elasticsearch.search.sort.SortOrder;

import com.alibaba.fastjson.JSON;

public class ElasticSearchTest {

  public static final String INDEX = "test";

  public static final String TYPE = "organization";

  public static TransportClient buildClient() throws UnknownHostException {

    return TransportClient.builder().build()

        .addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("localhost"), 9300));

  }

  public static void closeClient(TransportClient client) {

    if(client != null) {

      client.close();

    }

  }

  /**

   * 创建mapping（类似数据库的表结构）

   *

   * @param client

   *            es client

   * @param index

   *            索引名字

   * @param type

   *            索引下type的名字

   * @throws Exception

   */

  public static void createOrgMapping(TransportClient client) throws Exception {

    IndicesExistsResponse res = client.admin().indices().prepareExists(INDEX).execute().actionGet();

    if (res.isExists()) {

      DeleteIndexRequestBuilder delIdx = client.admin().indices().prepareDelete(INDEX);

      delIdx.execute().actionGet();

    }

    CreateIndexRequestBuilder createIndexRequestBuilder = client.admin().indices().prepareCreate(INDEX);

    final XContentBuilder organizationBuilder = XContentFactory.jsonBuilder()

        .startObject()

        .startObject(TYPE)

        .startObject("properties")

        .startObject("id")

        .field("type", "long")

        .endObject()

        .startObject("name")

        .field("type", "string")

        .endObject()

        .startObject("type")

        .field("type", "integer")

        .endObject()

        .startObject("industrialModeCodes")

        .field("type", "integer")

        .endObject()

        .startObject("updatedAt")

        .field("type", "date")

        .endObject()

        .endObject()

        .endObject()

        .endObject();

    createIndexRequestBuilder.addMapping(TYPE, organizationBuilder);

    // MAPPING DONE

    createIndexRequestBuilder.execute().actionGet();

  }

  /**

   * 往索引中添加数据

   *

   * @param client

   */

  public static void index(TransportClient client, ESOrganization organization) {

    IndexResponse response = client.prepareIndex(INDEX, TYPE, String.valueOf(organization.getId()))

        .setSource(JSON.toJSONString(organization)).get();

    System.out.println(response.getId());

  }

  public static void getById(TransportClient client, String id) {

    GetResponse response = client.prepareGet(INDEX, TYPE, id).get();

    System.out.println(response.getSourceAsString());

  }

  public static void deleteById(TransportClient client, String id) {

    DeleteResponse response = client.prepareDelete(INDEX, TYPE, id).get();

    System.out.println(response);

  }

  public static void update(TransportClient client, ESOrganization organization)

      throws InterruptedException, ExecutionException {

    UpdateRequest updateRequest = new UpdateRequest(INDEX, TYPE, String.valueOf(organization.getId()))

        .doc(JSON.toJSONString(organization));

    UpdateResponse response = client.update(updateRequest).get();

    System.out.println(response);

  }

  public static void multiGet(TransportClient client, String... ids) {

    MultiGetResponse multiGetItemResponses = client.prepareMultiGet().add(INDEX, TYPE, ids).get();

    for (MultiGetItemResponse itemResponse : multiGetItemResponses) {

      GetResponse response = itemResponse.getResponse();

      if (response.isExists()) {

        String json = response.getSourceAsString();

        System.out.println(json);

      }

    }

  }

  /**

   * 复杂一点的搜索

   * @param client

   * @param searchReq

   */

  public static void search(TransportClient client, OrgSearchReq searchReq) {

    BoolQueryBuilder orgQueryBuilder = QueryBuilders.boolQuery();

    if (searchReq.getKeyword() != null && searchReq.getKeyword().length > 0) {

      BoolQueryBuilder queryBuilder = QueryBuilders.boolQuery();

      for (String keyword : searchReq.getKeyword()) {

        queryBuilder.should(QueryBuilders.matchQuery("name", keyword));

      }

      orgQueryBuilder.must(queryBuilder);

    }

    if (!CollectionUtils.isEmpty(searchReq.getTypes())) {

      BoolQueryBuilder queryBuilder = QueryBuilders.boolQuery();

      for (Integer type : searchReq.getTypes()) {

        queryBuilder.should(QueryBuilders.matchQuery("type", type));

      }

      orgQueryBuilder.must(queryBuilder);

    }

    if (!CollectionUtils.isEmpty(searchReq.getIndustrialModeCodes())) {

        BoolQueryBuilder queryBuilder = QueryBuilders.boolQuery();

        for (Integer industrialMode : searchReq.getIndustrialModeCodes()) {

          queryBuilder.should(QueryBuilders.matchQuery("industrialModeCodes", industrialMode));

        }

        orgQueryBuilder.must(queryBuilder);

      }

    if (searchReq.getUpdatedAt() != null && searchReq.getUpdatedAt().length > 0) {

        if (searchReq.getUpdatedAt()[0] != null) {

          orgQueryBuilder.must(QueryBuilders.rangeQuery("updatedAt").from(searchReq.getUpdatedAt()[0]));

        }

        if (searchReq.getUpdatedAt().length == 2 && searchReq.getUpdatedAt()[1] != null) {

          orgQueryBuilder.must(QueryBuilders.rangeQuery("updatedAt").to(searchReq.getUpdatedAt()[1]));

        }

      }

    if (StringUtils.isEmpty(searchReq.getSortField())) {

      searchReq.setSortField("updatedAt");

    }

    System.out.println("orgQueryBuilder: " +orgQueryBuilder);

    SearchResponse searchResponse = client.prepareSearch(INDEX).setTypes(TYPE)

        .setQuery(orgQueryBuilder)

        .addSort(searchReq.getSortField(), getSortOrder(searchReq.getSortType()))

        .setFrom(searchReq.getOffset())

        .setSize(searchReq.getSize())

        .execute()

        .actionGet();

    List<ESOrganization> orgs = new ArrayList<>();

    for (SearchHit hit : searchResponse.getHits()) {

      ESOrganization org = JSON.parseObject(hit.getSourceAsString(), ESOrganization.class);

      orgs.add(org);

    }

    System.out.println(JSON.toJSONString(orgs, true));

  }

  private static SortOrder getSortOrder(String str) {

      if (StringUtils.isEmpty(str)) {

        return SortOrder.ASC;

      }

      for (SortOrder sortOrder : SortOrder.values()) {

        if (sortOrder.name().equalsIgnoreCase(str)) {

          return sortOrder;

        }

      }

      return SortOrder.ASC;

    }

  private static ESOrganization getUpdateOrg() {

    ESOrganization organization = new ESOrganization();

    organization.setId(1l);

    organization.setType(1111111111);

    return organization;

  }

  public static void addIndex(TransportClient client) {

    String[] orgNames = {

        "小米科技有限责任公司",

        "乐视网信息技术（北京）股份有限公司",

        "网易（杭州）网络有限公司",

        "阿里巴巴（中国）有限公司",

        "上海希格斯网络科技有限公司",

        "running man",

        "chicken run",

        "tom runs fast"

    };

    for (int i = 0; i< orgNames.length; i++) {

      ESOrganization organization = new ESOrganization();

      organization.setId(Long.valueOf(i+1));

      organization.setName(orgNames[i]);

      organization.setType(i * 1);

      organization.setIndustrialModeCodes(new Integer[]{1+i, 20+i, 30+i});

      organization.setUpdatedAt(new Date());

      index(client, organization);

    }

  }

   

  public static void main(String[] args) throws Exception {

    TransportClient client = null;

    try {

      client = buildClient();

      createOrgMapping(client);

      addIndex(client);

      update(client, getUpdateOrg());

      getById(client, "1");

      multiGet(client, "2", "3");

      deleteById(client, "8");

      //索引创建成功之后1秒钟左右，才会变得可用

      TimeUnit.SECONDS.sleep(1);

      search(client, getSearchReq());

    } finally {

      closeClient(client);

    }

     

     

     

  }

  private static OrgSearchReq getSearchReq() {

    OrgSearchReq req = new OrgSearchReq();

    req.setKeyword(new String[]{"阿里巴巴", "乐视"});

    req.setTypes(Arrays.asList(3));

    req.setIndustrialModeCodes(Arrays.asList(23,33));

    req.setUpdatedAt(new Date[]{null, new Date()});

    return req;

  }

}

package com.tch.test.elasticsearch.test;

import java.util.Date;

public class ESOrganization {

  private Long id;

  private String name;

  private Integer type;

  private Integer[] industrialModeCodes;

  private Date updatedAt;

  public Long getId() {

    return id;

  }

  public void setId(Long id) {

    this.id = id;

  }

  public String getName() {

    return name;

  }

  public void setName(String name) {

    this.name = name;

  }

  public Integer getType() {

    return type;

  }

  public void setType(Integer type) {

    this.type = type;

  }

  public Integer[] getIndustrialModeCodes() {

    return industrialModeCodes;

  }

  public void setIndustrialModeCodes(Integer[] industrialModeCodes) {

    this.industrialModeCodes = industrialModeCodes;

  }

  public Date getUpdatedAt() {

    return updatedAt;

  }

  public void setUpdatedAt(Date updatedAt) {

    this.updatedAt = updatedAt;

  }

}

package com.tch.test.elasticsearch.test;

import java.util.Date;

import java.util.List;

public class OrgSearchReq {

  private String[] keyword;

  private List<Integer> types;

  private List<Integer> industrialModeCodes;

  private Date[] updatedAt;

  private Integer page;

  private Integer size;

  private String sortField;

  private String sortType;

  private static final int FIRST_PAGE = 1;

  private static final int PAGE_SIZE = 10;

  public String[] getKeyword() {

    return keyword;

  }

  public void setKeyword(String[] keyword) {

    this.keyword = keyword;

  }

  public List<Integer> getTypes() {

    return types;

  }

  public void setTypes(List<Integer> types) {

    this.types = types;

  }

  public List<Integer> getIndustrialModeCodes() {

    return industrialModeCodes;

  }

  public void setIndustrialModeCodes(List<Integer> industrialModeCodes) {

    this.industrialModeCodes = industrialModeCodes;

  }

  public Date[] getUpdatedAt() {

    return updatedAt;

  }

  public void setUpdatedAt(Date[] updatedAt) {

    this.updatedAt = updatedAt;

  }

  public Integer getPage() {

    if (page == null || page < 0) {

      return FIRST_PAGE;

    }

    return page;

  }

  public void setPage(Integer page) {

    this.page = page;

  }

  public Integer getSize() {

    if (size == null || size < 0) {

      return PAGE_SIZE;

    }

    return size;

  }

  public void setSize(Integer size) {

    this.size = size;

  }

  public String getSortField() {

    return sortField;

  }

  public void setSortField(String sortField) {

    this.sortField = sortField;

  }

  public String getSortType() {

    return sortType;

  }

  public void setSortType(String sortType) {

    this.sortType = sortType;

  }

  public int getOffset() {

    return (getPage() - 1) * getSize();

  }

}

 

4.demo解释：

上面的demo中，ESOrganization.java为存到es中的数据结构，OrgSearchReq.java为搜索的条件包装类。

简单的CRUD就不说了，下面主要说一下搜索（search）的部分：

搜索中使用了

BoolQueryBuilder orgQueryBuilder = QueryBuilders.boolQuery();

在搜索条件复杂的情况下，我们只需要根据自己的需求使用 BoolQueryBuilder 的 must 和 should来添加 query 的条件，

这里的 must 类似 sql 中的 and， should 类似 sql 中的 or

例如上面的demo中：

BoolQueryBuilder orgQueryBuilder = QueryBuilders.boolQuery();

if (searchReq.getKeyword() != null && searchReq.getKeyword().length > 0) {

  BoolQueryBuilder queryBuilder = QueryBuilders.boolQuery();

  for (String keyword : searchReq.getKeyword()) {

    queryBuilder.should(QueryBuilders.matchQuery("name", keyword));

  }

  orgQueryBuilder.must(queryBuilder);

}

if (!CollectionUtils.isEmpty(searchReq.getTypes())) {

  BoolQueryBuilder queryBuilder = QueryBuilders.boolQuery();

  for (Integer type : searchReq.getTypes()) {

    queryBuilder.should(QueryBuilders.matchQuery("type", type));

  }

  orgQueryBuilder.must(queryBuilder);

}

if (!CollectionUtils.isEmpty(searchReq.getIndustrialModeCodes())) {

    BoolQueryBuilder queryBuilder = QueryBuilders.boolQuery();

    for (Integer industrialMode : searchReq.getIndustrialModeCodes()) {

      queryBuilder.should(QueryBuilders.matchQuery("industrialModeCodes", industrialMode));

    }

    orgQueryBuilder.must(queryBuilder);

  }

if (searchReq.getUpdatedAt() != null && searchReq.getUpdatedAt().length > 0) {

    if (searchReq.getUpdatedAt()[0] != null) {

      orgQueryBuilder.must(QueryBuilders.rangeQuery("updatedAt").from(searchReq.getUpdatedAt()[0]));

    }

    if (searchReq.getUpdatedAt().length == 2 && searchReq.getUpdatedAt()[1] != null) {

      orgQueryBuilder.must(QueryBuilders.rangeQuery("updatedAt").to(searchReq.getUpdatedAt()[1]));

    }

  }

if (StringUtils.isEmpty(searchReq.getSortField())) {

  searchReq.setSortField("updatedAt");

}

 

对应到sql语句就是：

where (name = 'xxx1' or name = 'xxx2') and (type = xxx3 or type = xxx4) and (industrialModeCodes = xxx5 or industrialModeCodes = xxx6) and updatedAt > xxx7 and updatedAt < xxx8

.addSort(searchReq.getSortField(), getSortOrder(searchReq.getSortType()))

.setFrom(searchReq.getOffset())

.setSize(searchReq.getSize())

对应的就是sql中的排序和分页。

但有一点不同就是

QueryBuilders.matchQuery("name", keyword)

和sql中的 name=xxx还有一点不同，因为前者会做分词，例如name为'小米有限公司' 经过分词之后，我们使用matchQuery搜索的时候，keyword输入 '小米' 也能搜到这个记录。

如果要完全等同，那就要使用：

QueryBuilders.termQuery("name", "小米")

termQuery不会使用分词，而会使用精确匹配。

范围搜索则是使用rangeQuery:

QueryBuilders.rangeQuery("updatedAt").from(xxx)

 

5.查看索引的mapping:

http://localhost:9200/test/organization/_mapping?pretty

 

6.查看索引中的记录：

http://localhost:9200/test/organization/_search?pretty

 

7.es中index和type的区别

https://www.elastic.co/blog/index-vs-type

 

 8.

2.0之后，filter已经和query合并：

https://www.elastic.co/blog/better-query-execution-coming-elasticsearch-2-0

 

9.elastic如何trouble shooting:

https://www.elastic.co/blog/found-beginner-troubleshooting