阅读量:0
easy-es、elasticsearch、分词器 与springboot 结合的代码我这里就不放了,我这里直接是使用代码。
基础准备:
创建实体类:
@Data // 索引名 @IndexName("test_jc") public class TestJcES { // id注解 @IndexId(type = IdType.CUSTOMIZE) private Long id; // 如果需要分词查询,必须 FieldType.TEXT analyzer = "ik_max_word" 官网有说明 @IndexField(fieldType = FieldType.TEXT, analyzer = "ik_max_word") private String name; // 非分词查询类型 最好用 KEYWORD @IndexField(fieldType = FieldType.KEYWORD) private String sex; /** * [描述] 如果某字段数组类型,并且该类型后期需要聚合操作,必须 fieldData = true * FieldType.TEXT:会将数组中的元素 “拆分单字符” 进行聚合 * FieldType.KEYWORD: 会对数组中的元素进行聚合 */ @IndexField(fieldType = FieldType.TEXT,fieldData = true) private List<String> industryTags; @IndexField(fieldType = FieldType.KEYWORD,fieldData = true) private List<String> productTags; //时间类型 @IndexField(fieldType = FieldType.DATE, dateFormat = "yyyy-MM-dd HH:mm:ss") private String updateTime; @IndexField(fieldType = FieldType.DATE, dateFormat = "yyyy-MM-dd HH:mm:ss") private String createTime; public TestJcES(Long id,String name, List<String> industryTags, List<String> productTags) { this.id = id; this.name = name; this.industryTags = industryTags; this.productTags = productTags; } }
PS:在easy-es的注解 @IndexFiled 中源码会有说明:
对应的mapper:
// BaseEsMapper 来自 easy-es框架 public interface TestJcESMapper extends cn.easyes.core.core.BaseEsMapper<TestJcES> { }
增删改(带批量):
testJcESMapper.deleteIndex("test_jd"); testJcESMapper.createIndex("test_jd"); TestJcES es = new TestJcES(1L,"小红",29,Arrays.asList("分类1","分类2","分类3"),Arrays.asList("标签1","标签2")); TestJcES es2 = new TestJcES(2L,"小白",29,Arrays.asList("分类1","分类3"),Arrays.asList("标签1","标签3")); TestJcES es3 = new TestJcES(3L,"小黑",30,Arrays.asList("分类4"),Arrays.asList("标签1")); TestJcES es4 = new TestJcES(4L,"小明",18,Arrays.asList("分类1"),Arrays.asList("标签1","标签2","变迁3")); testJcESMapper.insertBatch(Arrays.asList(es,es2,es3,es4)); //批量更新 //testJcESMapper.updateBatchByIds(Arrays.asList(es,es2,es3,es4)); //批量删除 //testJcESMapper.deleteBatchIds(Arrays.asList(1L,2L, 3L, 4L)); LambdaEsQueryWrapper<TestJcES> query = new LambdaEsQueryWrapper<>(); //相当于 select * from test_jc where name like '%红%' and sex = 29 and industryTags in ('标签1','标签2') query.and(item->item.match(TestJcES::getName, "红")); query.and(item->item.match(TestJcES::getSex, 29)); query.in("industryTags",Arrays.asList("标签1","标签2")); // 默认按查询度倒叙 lambdaEsQueryWrapper.sortByScore(SortOrder.DESC); //注意:从1开始起步 不是从0开始 EsPageInfo<TestJcES> pageQuery = testJcESMapper.pageQuery(query, 1, 10); //查询数据 System.out.println(pageQuery.getList()); //总条数 System.out.println(pageQuery.getTotal()); //总页数 System.out.println(pageQuery.getPages());
聚合操作:
1.普通keyword类型字段聚合:
LambdaEsQueryWrapper<TestJcES> query = new LambdaEsQueryWrapper<>(); //TODO 这里也可以通过query带条件进行聚合 //比如: query.match(TestJcES::getName, "红"); // 这里类似 select * from test_jc group by sex String filedName = "sex"; query.groupBy(filedName); // 是否统计hits的数据总数 设置为0 则不统计 数据量大的时候聚合速度会更快一些 //query.size(0); SearchResponse searchResponse = testJcESMapper.search(query); //7. 获取命中对象 SearchHits SearchHits hits = searchResponse.getHits(); //7.1 获取总记录数 如果 query.size(0) 则这里值就为0 Long total= hits.getTotalHits().value; System.out.println("被聚合的数据总条数:"+total); // aggregations 对象 Aggregations aggregations = searchResponse.getAggregations(); //将aggregations 转化为map Map<String, Aggregation> aggregationMap = aggregations.asMap(); //通过key获取 filedName+"Terms" 对象 使用Aggregation的子类接收 buckets属性在Terms接口中体现 // Aggregation goods_brands1 = aggregationMap.get(filedName+"Terms"); Terms resultTerms =(Terms) aggregationMap.get(filedName+"Terms"); //获取buckets 数组集合 List<? extends Terms.Bucket> buckets = resultTerms.getBuckets(); Map<String,Object>map=new HashMap<>(); //遍历buckets key 属性名,doc_count 统计聚合数 for (Terms.Bucket bucket : buckets) { System.out.println(bucket.getKey()); System.out.println(bucket.getDocCount()); map.put(bucket.getKeyAsString(),bucket.getDocCount()); }
聚合效果:
2.数组(text类型)类型聚合:
LambdaEsQueryWrapper<TestJcES> query = new LambdaEsQueryWrapper<>(); //TODO 这里也可以通过query带条件进行聚合 //比如: query.match(TestJcES::getName, "红"); String filedName = "industryTags"; query.groupBy(filedName); // 是否统计hits的数据总数 设置为0 则不统计 数据量大的时候聚合速度会更快一些 //query.size(0); SearchResponse searchResponse = testJcESMapper.search(query); //7. 获取命中对象 SearchHits SearchHits hits = searchResponse.getHits(); //7.1 获取总记录数 如果 query.size(0) 则这里值就为0 Long total= hits.getTotalHits().value; System.out.println("被聚合的数据总条数:"+total); // aggregations 对象 Aggregations aggregations = searchResponse.getAggregations(); //将aggregations 转化为map Map<String, Aggregation> aggregationMap = aggregations.asMap(); //通过key获取 filedName+"Terms" 对象 使用Aggregation的子类接收 buckets属性在Terms接口中体现 // Aggregation goods_brands1 = aggregationMap.get(filedName+"Terms"); Terms resultTerms =(Terms) aggregationMap.get(filedName+"Terms"); //获取buckets 数组集合 List<? extends Terms.Bucket> buckets = resultTerms.getBuckets(); Map<String,Object>map=new HashMap<>(); //遍历buckets key 属性名,doc_count 统计聚合数 for (Terms.Bucket bucket : buckets) { System.out.println(bucket.getKey()); System.out.println(bucket.getDocCount()); map.put(bucket.getKeyAsString(),bucket.getDocCount()); }
如果实体类的属性类型采用 text,则会把该属性里面的所有值分词然后进行聚合:
聚合效果:
2.数组(keyword类型)类型聚合:
LambdaEsQueryWrapper<TestJcES> query = new LambdaEsQueryWrapper<>(); //TODO 这里也可以通过query带条件进行聚合 //比如: query.match(TestJcES::getName, "红"); // 类似 select * from test_jc group by productTags String filedName = "productTags"; query.groupBy(filedName); // 是否统计hits的数据总数 设置为0 则不统计 数据量大的时候聚合速度会更快一些 //query.size(0); SearchResponse searchResponse = testJcESMapper.search(query); //7. 获取命中对象 SearchHits SearchHits hits = searchResponse.getHits(); //7.1 获取总记录数 如果 query.size(0) 则这里值就为0 Long total= hits.getTotalHits().value; System.out.println("被聚合的数据总条数:"+total); // aggregations 对象 Aggregations aggregations = searchResponse.getAggregations(); //将aggregations 转化为map Map<String, Aggregation> aggregationMap = aggregations.asMap(); //通过key获取 filedName+"Terms" 对象 使用Aggregation的子类接收 buckets属性在Terms接口中体现 // Aggregation goods_brands1 = aggregationMap.get(filedName+"Terms"); Terms resultTerms =(Terms) aggregationMap.get(filedName+"Terms"); //获取buckets 数组集合 List<? extends Terms.Bucket> buckets = resultTerms.getBuckets(); Map<String,Object>map=new HashMap<>(); //遍历buckets key 属性名,doc_count 统计聚合数 for (Terms.Bucket bucket : buckets) { System.out.println(bucket.getKey()); System.out.println(bucket.getDocCount()); map.put(bucket.getKeyAsString(),bucket.getDocCount()); }
聚合效果:
es聚合强大的地方在于,会把属性为数组拆分元素进行聚合统计,一般来说,普通统计用到这里就完全足够了。
PS 另外附赠elasticsearch通用聚合方法:
可
/** * [描述] */ private List<Map<String,Object>> commonGroup3(TestJcES search , String fieldName) { // 创建一个布尔查询来组合多个条件 BoolQueryBuilder boolQuery = QueryBuilders.boolQuery(); if (StringUtils.isNotBlank(search.getName())) { boolQuery.should(QueryBuilders.multiMatchQuery(search.getName(), "name")); } if(search.getProductTags() != null){ boolQuery.should(QueryBuilders.matchQuery("productTags", search.getProductTags())); } return commonGroupByBoolQuery(fieldName, boolQuery,"test_jc"); } /** * 根据布尔查询创建一个过滤聚合,并返回基于指定字段的聚合结果 * @param fieldName 指定的字段名 * @param boolQuery 基于该布尔查询创建过滤聚合 * @param indexName 索引名称 * @return 基于指定字段的聚合结果列表,每个结果包含字段名和计数 */ private List<Map<String, Object>> commonGroupByBoolQuery(String fieldName, BoolQueryBuilder boolQuery ,String indexName) { // 创建一个过滤聚合,基于布尔查询 FilterAggregationBuilder filterAgg = AggregationBuilders.filter("filtered_agg", boolQuery); // 在过滤后的文档上创建其他聚合 TermsAggregationBuilder termsAgg = AggregationBuilders.terms("agg_field") .field(fieldName); // 将聚合添加到过滤聚合中 filterAgg.subAggregation(termsAgg); SearchRequest searchRequest = new SearchRequest(indexName); SearchSourceBuilder sourceBuilder = new SearchSourceBuilder(); // 添加聚合到搜索源构建器 sourceBuilder.aggregation(filterAgg); searchRequest.source(sourceBuilder); try { SearchResponse searchResponse = restHighLevelClient.search(searchRequest, RequestOptions.DEFAULT); // 获取聚合结果 Filter filteredAggregation = searchResponse.getAggregations().get("filtered_agg"); Terms yourFieldAggregation = filteredAggregation.getAggregations().get("agg_field"); return yourFieldAggregation.getBuckets().stream() .map(item -> { Map<String, Object> map = new HashMap<>(2); map.put("name", item.getKeyAsString()); map.put("count", item.getDocCount()); return map; }) .collect(Collectors.toList()); } catch (IOException e) { e.printStackTrace(); } return List.of(); }
另附easy-es官网地址:
https://www.easy-es.cn/pages/ce1922/#%E5%B8%B8%E8%A7%84%E8%81%9A%E5%90%88
官网很完整的demo:
https://www.easy-es.cn/pages/17ea0a/#%E4%BC%98%E5%8A%BF%E5%AF%B9%E6%AF%94
部分es教程博客:
https://blog.csdn.net/weixin_46115287/article/details/120974337