这篇文章主要介绍“Grouping组查询怎么使用”,在日常操作中,相信很多人在Grouping组查询怎么使用问题上存在疑惑,小编查阅了各式资料,整理出简单好用的操作方法,希望对大家解答”Grouping组查询怎么使用”的疑惑有所帮助!接下来,请跟着小编一起来学习吧!
lucene的联想词是在org.apache.lucene.lucene-grouping包下边,提供了组查询功能的支持。
Grouping介绍
Grouping主要为用户提供了处理不同lucene的中含有相同filed的不同doc的分组统计。
<!--组查询--> <dependency> <groupId>org.apache.lucene</groupId> <artifactId>lucene-grouping</artifactId> <version>5.5.2</version> </dependency>
Grouping深入
1. Grouping参数
groupField:要分组的字段(如果groupField在doc中不存在,会返回一个null的分组) groupSort:分组的排序规则,排序字段决定了分组内容展示的先后顺序 topNGroup:分组展示的数量,只计算0到topNGroup条记录 groupOffset:从第几个TopGroup开始计算 例:groupOffset为3的话,会展示从3到topNGroup对应的记录,此数值我们可以用于分页查询 withinGroupSort:每组内怎么排序 maxDocsPerGroup:每组处理多少个doc withinGroupOffset:每组显示的doc初始位置
2.Grouping实现
grouping实现需要两步:
-
利用TermFirstPassGroupingCollector来收集top groups
-
用TermSecondPassGroupingCollector处理每个group对应的doc
3. Grouping查询
对要搜索的信息创建Query查询对象,Lucene会根据Query查询对象生成最终的查询语法,类似关系数据库Sql语法一样Lucene也有自己的查询语法,比如:"word:北京长安中西医结合医院"表示查询Field的word为"北京长安中西医结合医院"的文档信息。
可通过两种方法创建查询对象:
-
使用Lucene提供Query子类
Query是一个抽象类,lucene提供了很多查询对象,比如TermQuery项精确查询,NumericRangeQuery数字范围查询等。
Query query = new TermQuery(new Term("word", "北京市鼓楼中医院")); //参数1:要查询的域 参数2:最小值 参数3:最大值 参数4:是否包含最小值 参数5:是否包含最大值 Query query =NumericRangeQuery.newLongRange("size", 0, 1000, true, true); 扩展: BooleanQuery query = new BooleanQuery(); Query query1 =NumericRangeQuery.newLongRange("size", 0, 1000, true, true); Query query2 = new TermQuery(new Term("word","北京市鼓楼中医院")); //添加到BooleanQuery对象中 query.add(query1,Occur.MUST); query.add(query2,Occur.MUST); 备注: Occur.MUST:必须满足此条件 Occur.SHOULD:应该满足,但是不满足也可以 Occur.MUST_NOT:必须不满足
-
使用QueryParse解析查询表达式
QueryParse会将用户输入的查询表达式解析成Query对象实例。
QueryParser queryParser =new QueryParser("word", new StandardAnalyzer()); Query query = queryParser.parse("北京市鼓楼中医院");
Grouping用例
1. Controller层
package com.spring.master.lucene.group.controller; import com.spring.master.lucene.group.service.GroupService; import org.springframework.web.bind.annotation.GetMapping; import org.springframework.web.bind.annotation.RequestMapping; import org.springframework.web.bind.annotation.RestController; import javax.annotation.Resource; /** * @author Huan Lee * @version 1.0 * @date 2020-09-14 14:19 * @describtion 业精于勤,荒于嬉;行成于思,毁于随。 */ @RestController @RequestMapping(value = "/group") public class GroupController { @Resource private GroupService groupService; /** * 创建索引文档 * @return */ @GetMapping(value = "/createIndex") public String createIndex() { return groupService.createIndex(); } /** * 搜索分组 * lucene 5 * @return */ @GetMapping(value = "/searchGroup5") public String searchGroup5(String keyword) { groupService.searchGroup5(keyword); return "true"; } /** * 搜索分组 * lucene 7 * @return */ @GetMapping(value = "/searchGroup7") public String searchGroup7(String keyword) { groupService.searchGroup7(keyword); return "true"; } } 访问地址: localhost:2000/spring-master/group/createIndex localhost:2000/spring-master/group/searchGroup5?keyword=北京长安中西医结合医院 localhost:2000/spring-master/group/searchGroup7?keyword=北京长安中西医结合医院
2. Service层
package com.spring.master.lucene.group.impl; import com.spring.master.global.Identities; import com.spring.master.lucene.group.constant.SystemConstants; import com.spring.master.lucene.group.service.GroupService; import com.spring.master.lucene.suggest.constant.SuggestConstants; import com.spring.master.lucene.suggest.util.FileUtils; import com.spring.master.lucene.suggest.vo.DictionaryVO; import lombok.extern.slf4j.Slf4j; import org.apache.lucene.analysis.Analyzer; import org.apache.lucene.analysis.standard.StandardAnalyzer; import org.apache.lucene.document.*; import org.apache.lucene.index.*; import org.apache.lucene.queryparser.classic.MultiFieldQueryParser; import org.apache.lucene.queryparser.classic.QueryParser; import org.apache.lucene.queryparser.xml.builders.MatchAllDocsQueryBuilder; import org.apache.lucene.search.*; import org.apache.lucene.search.grouping.GroupDocs; import org.apache.lucene.search.grouping.GroupingSearch; import org.apache.lucene.search.grouping.SearchGroup; import org.apache.lucene.search.grouping.TopGroups; import org.apache.lucene.search.grouping.term.TermAllGroupsCollector; import org.apache.lucene.search.grouping.term.TermFirstPassGroupingCollector; import org.apache.lucene.search.grouping.term.TermSecondPassGroupingCollector; import org.apache.lucene.store.Directory; import org.apache.lucene.store.FSDirectory; import org.apache.lucene.util.BytesRef; import org.apache.lucene.util.Version; import org.nlpcn.commons.lang.util.MD5; import org.springframework.stereotype.Service; import java.io.IOException; import java.nio.file.Paths; import java.util.ArrayList; import java.util.Collection; import java.util.List; import java.util.Random; /** * @author Huan Lee * @version 1.0 * @date 2020-09-14 18:45 * @describtion 业精于勤,荒于嬉;行成于思,毁于随。 */ @Service @Slf4j public class GroupServiceImpl implements GroupService { /** * 指定在哪个索引上进行分组 */ static String groupField = "sourceType"; /** * 标准分词器 */ private static Analyzer analyzer = new StandardAnalyzer(); @Override public String createIndex() { try { Directory directory = FSDirectory.open(Paths.get(SystemConstants.indexDir)); IndexWriterConfig indexWriterConfig = new IndexWriterConfig(analyzer); indexWriterConfig.setOpenMode(IndexWriterConfig.OpenMode.CREATE_OR_APPEND); IndexWriter writer = new IndexWriter(directory, indexWriterConfig); // 读DictionaryVO数据 List<DictionaryVO> diseases = FileUtils.readCsv(SuggestConstants.disease); List<DictionaryVO> doctors = FileUtils.readCsv(SuggestConstants.doctor); List<DictionaryVO> facultys = FileUtils.readCsv(SuggestConstants.faculty); List<DictionaryVO> hospitals = FileUtils.readCsv(SuggestConstants.hospital); List<DictionaryVO> drugcatalogues = FileUtils.readCsv(SuggestConstants.drugcatalogue); diseases.forEach(disease -> { Document doc = new Document(); // 进行分组的域上建立的必须是SortedDocValuesField类型 doc.add(new SortedDocValuesField(groupField, new BytesRef("Disease"))); doc.add(new StringField("id", Identities.uuid(), Field.Store.YES)); doc.add(new StringField("sourceType", "Disease", Field.Store.YES)); doc.add(new TextField("word", disease.getWord(), Field.Store.YES)); try { writer.addDocument(doc); } catch (Exception e) { log.error(e.getMessage()); } }); hospitals.forEach(hospital -> { Document doc = new Document(); // 进行分组的域上建立的必须是SortedDocValuesField类型 doc.add(new SortedDocValuesField(groupField, new BytesRef("Hospital"))); doc.add(new StringField("id", Identities.uuid(), Field.Store.YES)); doc.add(new StringField("sourceType", "Hospital", Field.Store.YES)); doc.add(new TextField("word", hospital.getWord(), Field.Store.YES)); try { writer.addDocument(doc); } catch (Exception e) { log.error(e.getMessage()); } }); facultys.forEach(faculty -> { Document doc = new Document(); // 进行分组的域上建立的必须是SortedDocValuesField类型 doc.add(new SortedDocValuesField(groupField, new BytesRef("Faculty"))); doc.add(new StringField("id", Identities.uuid(), Field.Store.YES)); doc.add(new StringField("sourceType", "Faculty", Field.Store.YES)); doc.add(new TextField("word", faculty.getWord(), Field.Store.YES)); try { writer.addDocument(doc); } catch (Exception e) { log.error(e.getMessage()); } }); drugcatalogues.forEach(drugcatalogue -> { Document doc = new Document(); // 进行分组的域上建立的必须是SortedDocValuesField类型 doc.add(new SortedDocValuesField(groupField, new BytesRef("Drugcatalogue"))); doc.add(new StringField("id", Identities.uuid(), Field.Store.YES)); doc.add(new StringField("sourceType", "Drugcatalogue", Field.Store.YES)); doc.add(new TextField("word", drugcatalogue.getWord(), Field.Store.YES)); try { writer.addDocument(doc); } catch (Exception e) { log.error(e.getMessage()); } }); doctors.forEach(doctor -> { Document doc = new Document(); // 进行分组的域上建立的必须是SortedDocValuesField类型 doc.add(new SortedDocValuesField(groupField, new BytesRef("Doctor"))); doc.add(new StringField("id", Identities.uuid(), Field.Store.YES)); doc.add(new StringField("sourceType", "Doctor", Field.Store.YES)); doc.add(new TextField("word", doctor.getWord(), Field.Store.YES)); try { writer.addDocument(doc); } catch (Exception e) { log.error(e.getMessage()); } }); // 6 -- no sourceType field Document doc = new Document(); doc.add(new StringField("sourceType", "Faculty", Field.Store.YES)); doc.add(new TextField("word", "高压氧室", Field.Store.YES)); doc.add(new StringField("id", "0", Field.Store.YES)); writer.addDocument(doc); writer.commit(); writer.close(); } catch (Exception e) { log.error(e.getMessage()); } return "true"; } @Override public void searchGroup5(String keyword) { try { Directory directory = FSDirectory.open(Paths.get(SystemConstants.indexDir)); IndexReader reader = DirectoryReader.open(directory); IndexSearcher searcher = new IndexSearcher(reader); // term查询 // Query query = new TermQuery(new Term("word", keyword)); QueryParser queryParser =new QueryParser("word", new StandardAnalyzer()); Query query = queryParser.parse(keyword); // 每个分组内部的排序规则 Sort groupSort = Sort.RELEVANCE; // 前N条中分组 int topNGroups = 10; // 分组起始偏移量 int groupOffset = 0; // 是否填充SearchGroup的sortValues boolean fillFields = true; // groupSort用于对组进行排序,docSort用于对组内记录进行排序,多数情况下两者是相同的,但也可不同 Sort docSort = groupSort; // 用于组内分页,起始偏移量 int docOffset = 0; // 每组返回多少条结果 int docsPerGroup = 5; // 是否需要计算总的分组数量 boolean requiredTotalGroupCount = true; // 是否需要缓存评分 boolean cacheScores = true; // 第一次查询缓存容量的大小:设置为16M double maxCacheRAMMB = 16.0; // 支持lucene5.x版本 TermFirstPassGroupingCollector c1 = new TermFirstPassGroupingCollector("sourceType", groupSort, groupOffset + topNGroups); /** 将TermFirstPassGroupingCollector包装成CachingCollector,为第一次查询加缓存,避免重复评分 * CachingCollector就是用来为结果收集器添加缓存功能的 */ CachingCollector cachedCollector = CachingCollector.create(c1, cacheScores, maxCacheRAMMB); // 开始第一次分组统计 searcher.search(query, cachedCollector); // 第一次查询返回的结果集TopGroups中只有分组域值以及每组总的评分,至于每个分组里有几条,分别哪些索引文档,则需要进行第二次查询获取 Collection<SearchGroup<BytesRef>> topGroups = c1.getTopGroups(groupOffset, fillFields); if (topGroups == null) { System.out.println("No groups matched "); return; } // 是否获取每个分组内部每个索引的评分 boolean getScores = true; // 是否计算最大评分 boolean getMaxScores = true; // 如果需要对Lucene的score进行修正,则需要重载TermSecondPassGroupingCollector TermSecondPassGroupingCollector c2 = new TermSecondPassGroupingCollector("sourceType", topGroups, groupSort, docSort, docOffset + docsPerGroup, getScores, getMaxScores, fillFields); // 如果需要计算总的分组数量,则需要把TermSecondPassGroupingCollector包装成TermAllGroupsCollector // TermAllGroupsCollector就是用来收集总分组数量的 TermAllGroupsCollector allGroupsCollector = null; Collector secondPassCollector = null; //若需要统计总的分组数量 if (requiredTotalGroupCount) { allGroupsCollector = new TermAllGroupsCollector("sourceType"); secondPassCollector = MultiCollector.wrap(c2, allGroupsCollector); } else { secondPassCollector = c2; } // 如果第一次查询已经加了缓存,则直接从缓存中取 if (cachedCollector.isCached()) { // 第二次查询直接从缓存中取 cachedCollector.replay(secondPassCollector); } else { // 开始第二次分组查询 searcher.search(query, secondPassCollector); } // 所有组的数量 int totalGroupCount = 0; // 所有满足条件的记录数 int totalHitCount = 0; // 所有组内的满足条件的记录数(通常该值与totalHitCount是一致的) int totalGroupedHitCount = -1; if (requiredTotalGroupCount) { totalGroupCount = allGroupsCollector.getGroupCount(); } TopGroups<BytesRef> groupsResult = c2.getTopGroups(docOffset); // 这里打印的3项信息就是第一次查询的统计结果 totalHitCount = groupsResult.totalHitCount; totalGroupedHitCount = groupsResult.totalGroupedHitCount; // 打印总的分组数量 log.info("groupCount: {}", totalGroupCount); log.info("groupsResult.totalHitCount: {}", totalHitCount); log.info("groupsResult.totalGroupedHitCount: {}", totalGroupedHitCount); log.info("************************************"); int groupIdx = 0; // 下面打印的是第二次查询的统计结果,如果你仅仅只需要第一次查询的统计结果信息,不需要每个分组内部的详细信息,则不需要进行第二次查询,请知晓 // 迭代组 for (GroupDocs<BytesRef> groupDocs : groupsResult.groups) { groupIdx ++; String groupVL = groupDocs.groupValue == null ? "分组域的域值为空" : new String(groupDocs.groupValue.bytes); // 分组域的域值,groupIdx表示组的索引即第几组 log.info("group[{}].groupFieldValue: {}", groupIdx, groupVL); // 当前分组内命中的总记录数 log.info("group[{}].totalHits: {}", groupIdx, groupDocs.totalHits); int docIdx = 0; // 迭代组内的记录 for (ScoreDoc scoreDoc : groupDocs.scoreDocs) { docIdx++; // 打印分组内部每条记录的索引文档ID及其评分 log.info("group[" + groupIdx + "][" + docIdx + "]{docID:Score}:" + scoreDoc.doc + "/" + scoreDoc.score); // 根据docID可以获取到整个Document对象,通过doc.get(fieldName)可以获取某个存储域的域值 // 注意searcher.doc根据docID返回的document对象中不包含docValuesField域的域值,只包含非docValuesField域的域值,请知晓 Document doc = searcher.doc(scoreDoc.doc); log.info("group[" + groupIdx + "][" + docIdx + "]{docID:author}:" + doc.get("id") + ":" + doc.get("word")); } log.info("*****************************************"); } } catch (Exception e) { log.error(e.getMessage()); } } @Override public void searchGroup7(String keyword) { try { Directory directory = FSDirectory.open(Paths.get(SystemConstants.indexDir)); IndexReader reader = DirectoryReader.open(directory); IndexSearcher searcher = new IndexSearcher(reader); // 标准分词 Analyzer analyzer = new StandardAnalyzer(); // 指定要进行分组的索引 GroupingSearch groupingSearch = new GroupingSearch(groupField); // 指定分组排序规则 groupingSearch.setGroupSort(new Sort(SortField.FIELD_SCORE)); // 是否填充SearchGroup的sortValues groupingSearch.setFillSortFields(true); groupingSearch.setCachingInMB(4.0, true); groupingSearch.setAllGroups(true); // groupingSearch.setAllGroupHeads(true); // 限制分组个数 groupingSearch.setGroupDocsLimit(10); QueryParser parser = new QueryParser("word", analyzer); Query query = parser.parse(keyword); TopGroups<BytesRef> result = groupingSearch.search(searcher, query, 0, 1000); // 总命中数 log.info("总命中数: {}", result.totalHitCount); // log.info("分组数:{}", result.groups.length); // 按照分组打印查询结果 for (GroupDocs<BytesRef> groupDocs : result.groups){ if (groupDocs != null) { if (groupDocs.groupValue != null) { log.info("分组:{}", groupDocs.groupValue.utf8ToString()); }else{ // 由于建立索引时有一条数据没有在分组索引上建立SortedDocValued索引,因此这个分组的groupValue为null log.info("分组:{}", "unknow"); } log.info("组内数据条数:{}", groupDocs.totalHits); for(ScoreDoc scoreDoc : groupDocs.scoreDocs){ log.info("sourceType:{}", searcher.doc(scoreDoc.doc).get("sourceType")); log.info("word:{}", searcher.doc(scoreDoc.doc).get("word")); log.info("*****************************"); } System.out.println("====================================="); } } } catch (Exception e) { log.error(e.getMessage()); } } }
3. Util
package com.spring.master.lucene.util; import com.spring.master.lucene.suggest.vo.DictionaryVO; import lombok.extern.slf4j.Slf4j; import java.io.BufferedReader; import java.io.FileReader; import java.util.ArrayList; import java.util.List; /** * @author Huan Lee * @version 1.0 * @date 2020-09-11 09:57 * @describtion 业精于勤,荒于嬉;行成于思,毁于随。 */ @Slf4j public class FileUtils { /** * 读取词典csv文件 * @param fileNamePath * @return */ public static List<DictionaryVO> readCsv(String fileNamePath) { List<DictionaryVO> dictionarys = new ArrayList<>(); try { // 换成你的文件名 BufferedReader reader = new BufferedReader(new FileReader(fileNamePath)); String line; while ((line = reader.readLine()) != null) { // CSV格式文件为逗号分隔符文件,这里根据逗号切分 String[] item = line.split(","); dictionarys.add(new DictionaryVO(item[0], item[1], Long.parseLong(item[2]), Long.parseLong(item[3]))); } } catch (Exception e) { e.printStackTrace(); log.error(e.getMessage()); } return dictionarys; } }
4. Constant
package com.spring.master.lucene.group.constant; /** * @author Huan Lee * @version 1.0 * @date 2020-09-14 14:27 * @describtion 业精于勤,荒于嬉;行成于思,毁于随。 */ public class SystemConstants { /** * 索引目录 */ public static final String indexDir = "/Users/lihuan/Documents/projects/git/me/dictionary"; }
到此,关于“Grouping组查询怎么使用”的学习就结束了,希望能够解决大家的疑惑。理论与实践的搭配能更好的帮助大家学习,快去试试吧!若想继续学习更多相关知识,请继续关注亿速云网站,小编会继续努力为大家带来更多实用的文章!
原创文章,作者:745907710,如若转载,请注明出处:https://blog.ytso.com/223996.html