| /*------------------------------------------------------------------------------ |
| * Copyright (C) 2003-2006 Ben van Klinken and the CLucene Team |
| * |
| * Distributable under the terms of either the Apache License (Version 2.0) or |
| * the GNU Lesser General Public License, as specified in the COPYING file. |
| ------------------------------------------------------------------------------*/ |
| #include <memory> |
| #include "test.h" |
| #include "CLucene/analysis/cjk/CJKAnalyzer.h" |
| #include "CLucene/analysis/LanguageBasedAnalyzer.h" |
| #ifdef _CL_HAVE_IO_H |
| #include <io.h> |
| #endif |
| #ifdef _CL_HAVE_SYS_STAT_H |
| #include <sys/stat.h> |
| #endif |
| #ifdef _CL_HAVE_UNISTD_H |
| #include <unistd.h> |
| #endif |
| #ifdef _CL_HAVE_DIRECT_H |
| #include <direct.h> |
| #endif |
| |
| CL_NS_USE2(analysis, cjk) |
| |
| void test(CuTest* tc, char* orig, Reader* reader, bool verbose, int64_t bytes) { |
| StandardAnalyzer analyzer; |
| TokenStream* stream = analyzer.tokenStream(NULL, reader); |
| |
| uint64_t start = Misc::currentTimeMillis(); |
| |
| int32_t count = 0; |
| Token t; |
| char atmp[LUCENE_MAX_WORD_LEN + 1]; |
| TCHAR ttmp[LUCENE_MAX_WORD_LEN + 1]; |
| for (; stream->next(&t);) { |
| if (verbose) { |
| CuMessage(tc, _T("Text=%s start=%d end=%d\n"), t.termBuffer<TCHAR>(), t.startOffset(), |
| t.endOffset()); |
| } |
| int len = t.termLength<TCHAR>(); |
| |
| //use the lucene strlwr function (so copy to TCHAR first then back) |
| strncpy(atmp, orig + t.startOffset(), len); |
| atmp[len] = 0; |
| STRCPY_AtoT(ttmp, atmp, len + 1); |
| _tcslwr(ttmp); |
| |
| if (_tcsncmp(t.termBuffer<TCHAR>(), ttmp, len) != 0) { |
| TCHAR err[1024]; |
| _sntprintf(err, 1024, _T("token '%s' didnt match original text at %d-%d"), |
| t.termBuffer<TCHAR>(), t.startOffset(), t.endOffset()); |
| CuAssert(tc, err, false); |
| } |
| |
| // _CLDELETE(t); |
| count++; |
| } |
| |
| uint64_t end = Misc::currentTimeMillis(); |
| int64_t time = end - start; |
| CuMessageA(tc, "%d milliseconds to extract ", time); |
| CuMessageA(tc, "%d tokens\n", count); |
| CuMessageA(tc, "%f microseconds/token\n", (time * 1000.0) / count); |
| CuMessageA(tc, "%f megabytes/hour\n", (bytes * 1000.0 * 60.0 * 60.0) / (time * 1000000.0)); |
| |
| _CLDELETE(stream); |
| } |
| |
| void _testFile(CuTest* tc, const char* fname, bool verbose) { |
| struct fileStat buf; |
| fileStat(fname, &buf); |
| int64_t bytes = buf.st_size; |
| |
| char* orig = _CL_NEWARRAY(char, bytes); |
| { |
| FILE* f = fopen(fname, "rb"); |
| int64_t r = fread(orig, bytes, 1, f); |
| fclose(f); |
| } |
| |
| CuMessageA(tc, " Reading test file containing %d bytes.\n", bytes); |
| jstreams::FileReader fr(fname, "ASCII"); |
| const TCHAR* start; |
| size_t total = 0; |
| int32_t numRead; |
| do { |
| numRead = fr.read((const void**)&start, 1, 0); |
| if (numRead == -1) break; |
| total += numRead; |
| } while (numRead >= 0); |
| |
| jstreams::FileReader reader(fname, "ASCII"); |
| |
| test(tc, orig, &reader, verbose, total); |
| |
| _CLDELETE_CaARRAY(orig); |
| } |
| |
| void testFile(CuTest* tc) { |
| char loc[1024]; |
| strcpy(loc, clucene_data_location); |
| strcat(loc, "/reuters-21578/feldman-cia-worldfactbook-data.txt"); |
| CuAssert(tc, _T("reuters-21578/feldman-cia-worldfactbook-data.txt does not exist"), |
| Misc::dir_Exists(loc)); |
| _testFile(tc, loc, false); |
| } |
| |
| void _testCJK(CuTest* tc, const char* astr, const char** results, bool ignoreSurrogates = true) { |
| //SimpleInputStreamReader r(new AStringReader(astr), SimpleInputStreamReader::UTF8); |
| auto r = std::make_unique<lucene::util::SStringReader<char>>(astr, strlen(astr), false); |
| |
| CJKTokenizer* tokenizer = _CLNEW CJKTokenizer(r.get()); |
| tokenizer->setIgnoreSurrogates(ignoreSurrogates); |
| int pos = 0; |
| Token tok; |
| //TCHAR tres[LUCENE_MAX_WORD_LEN]; |
| |
| while (results[pos] != NULL) { |
| CLUCENE_ASSERT(tokenizer->next(&tok) != NULL); |
| |
| //lucene_utf8towcs(tres, results[pos], LUCENE_MAX_WORD_LEN); |
| //cout << results[pos] << " actual " << std::string(tok.termBuffer<char>() ,tok.termLength<char>())<< std::endl; |
| CLUCENE_ASSERT(strncmp(tok.termBuffer<char>(), results[pos], tok.termLength<char>()) == 0); |
| //CuAssertStrEquals(tc, "unexpected token value", tres, tok.termBuffer<char>()); |
| pos++; |
| } |
| CLUCENE_ASSERT(!tokenizer->next(&tok)); |
| |
| _CLDELETE(tokenizer); |
| } |
| |
| void testCJK(CuTest* tc) { |
| //utf16 test |
| //we have a very large unicode character: |
| //xEFFFF = utf8(F3 AF BF BF) = utf16(DB7F DFFF) = utf8(ED AD BF, ED BF BF) |
| static const char* exp4[4] = {"我爱", "爱你", "", NULL}; |
| _testCJK(tc, "我爱你", exp4, false); |
| |
| static const char* exp3[4] = {"\xED\xAD\xBF\xED\xBF\xBF\xe5\x95\xa4", |
| "\xe5\x95\xa4\xED\xAD\xBF\xED\xBF\xBF", "", NULL}; |
| _testCJK(tc, "\xED\xAD\xBF\xED\xBF\xBF\xe5\x95\xa4\xED\xAD\xBF\xED\xBF\xBF", exp3, false); |
| |
| static const char* exp1[5] = {"test", "t\xc3\xbcrm", "values", NULL}; |
| _testCJK(tc, "test t\xc3\xbcrm values", exp1); |
| |
| static const char* exp2[6] = { |
| "a", "\xe5\x95\xa4\xe9\x85\x92", "\xe9\x85\x92\xe5\x95\xa4", "", "x", NULL}; |
| _testCJK(tc, "a\xe5\x95\xa4\xe9\x85\x92\xe5\x95\xa4x", exp2); |
| } |
| |
| void testSimpleJiebaSearchModeTokenizer2(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "冰咒龙"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Search); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "冰咒", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "龙", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaAllModeTokenizer2(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "冰咒龙"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::All); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "冰", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "咒", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "龙", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaAllModeTokenizer(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "我来到北京清华大学"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::All); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "我", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "来到", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "北京", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "清华", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "清华大学", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "华大", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "大学", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaDefaultModeTokenizer2(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "中国的科技发展在世界上处于领先"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Default); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| /*char tmp[255] = {}; |
| while(ts->next(&t) != nullptr) { |
| lucene_wcstoutf8(tmp, t.termBuffer<TCHAR>(), 254); |
| std::cout << tmp << std::endl; |
| }*/ |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "中国", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "科技", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "发展", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "在世界上", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "处于", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "领先", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaDefaultModeTokenizer(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "我来到北京清华大学"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Default); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "我", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "来到", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "北京", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "清华大学", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaSearchModeTokenizer(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "我来到北京清华大学"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Search); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "我", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "来到", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "北京", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "清华", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "华大", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "大学", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "清华大学", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaTokenizer(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "我爱你中国"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Default); |
| a.initDict("./dict"); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "我爱你", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "中国", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaTokenizer2(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "人民可以得到更多实惠"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Default); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "人民", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "可以", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "得到", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "更", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "多", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "实惠", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaTokenizer3(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "中国人民银行"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| a.setMode(lucene::analysis::AnalyzerMode::Default); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "中国人民银行", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testSimpleJiebaTokenizer4(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| const char* field_value_data = "人民，银行"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| TokenStream* ts; |
| Token t; |
| |
| //test with chinese |
| a.setLanguage(_T("chinese")); |
| a.setStem(false); |
| ts = a.tokenStream(_T("contents"), stringReader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "人民", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "银行", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) == NULL); |
| _CLDELETE(ts); |
| } |
| |
| void testChineseAnalyzer(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| //CL_NS(util)::StringReader reader(_T("我爱你")); |
| auto reader = |
| std::make_unique<lucene::util::SStringReader<char>>("我爱你", strlen("我爱你"), false); |
| //reader->mark(50); |
| TokenStream* ts; |
| Token t; |
| |
| //test with cjk |
| a.setLanguage(_T("cjk")); |
| a.setStem(false); |
| ts = a.tokenStream(_T("contents"), reader.get()); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "我爱", t.termLength<char>()) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(strncmp(t.termBuffer<char>(), "爱你", t.termLength<char>()) == 0); |
| _CLDELETE(ts); |
| } |
| |
| void testChinese(CuTest* tc) { |
| RAMDirectory dir; |
| |
| auto analyzer = _CLNEW lucene::analysis::LanguageBasedAnalyzer(); |
| analyzer->setLanguage(L"cjk"); |
| |
| IndexWriter w(&dir, analyzer, true); |
| auto field_name = lucene::util::Misc::_charToWide("chinese"); |
| |
| Document doc; |
| auto field = _CLNEW Field(field_name, Field::INDEX_TOKENIZED | Field::STORE_NO); |
| doc.add(*field); |
| |
| const char* field_value_data = "人民可以得到更多实惠"; |
| auto stringReader = _CLNEW lucene::util::SimpleInputStreamReader( |
| new lucene::util::AStringReader(field_value_data), |
| lucene::util::SimpleInputStreamReader::UTF8); |
| field->setValue(stringReader); |
| w.addDocument(&doc); |
| |
| const char* field_value_data1 = "中国人民银行"; |
| auto stringReader1 = _CLNEW lucene::util::SimpleInputStreamReader( |
| new lucene::util::AStringReader(field_value_data1), |
| lucene::util::SimpleInputStreamReader::UTF8); |
| field->setValue(stringReader1); |
| w.addDocument(&doc); |
| |
| const char* field_value_data2 = "洛杉矶人，洛杉矶居民"; |
| auto stringReader2 = _CLNEW lucene::util::SimpleInputStreamReader( |
| new lucene::util::AStringReader(field_value_data2), |
| lucene::util::SimpleInputStreamReader::UTF8); |
| field->setValue(stringReader2); |
| w.addDocument(&doc); |
| |
| const char* field_value_data3 = "民族，人民"; |
| auto stringReader3 = _CLNEW lucene::util::SimpleInputStreamReader( |
| new lucene::util::AStringReader(field_value_data3), |
| lucene::util::SimpleInputStreamReader::UTF8); |
| field->setValue(stringReader3); |
| w.addDocument(&doc); |
| |
| w.close(); |
| |
| IndexSearcher searcher(&dir); |
| Term* t1 = _CLNEW Term(_T("chinese"), _T("人民")); |
| auto* query1 = _CLNEW TermQuery(t1); |
| Hits* hits1 = searcher.search(query1); |
| CLUCENE_ASSERT(3 == hits1->length()); |
| Term* t2 = _CLNEW Term(_T("chinese"), _T("民族")); |
| auto* query2 = _CLNEW TermQuery(t2); |
| Hits* hits2 = searcher.search(query2); |
| CLUCENE_ASSERT(1 == hits2->length()); |
| |
| doc.clear(); |
| //_CLDELETE(field) |
| _CLDELETE(hits1) |
| _CLDELETE(hits2) |
| } |
| |
| void testJiebaMatch(CuTest* tc) { |
| RAMDirectory dir; |
| auto field_name = lucene::util::Misc::_charToWide("chinese"); |
| try { |
| auto analyzer = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(); |
| analyzer->setLanguage(L"chinese"); |
| analyzer->setMode(lucene::analysis::AnalyzerMode::Default); |
| IndexWriter w(&dir, analyzer.get(), true); |
| w.setUseCompoundFile(false); |
| |
| Document doc; |
| auto field = _CLNEW Field(field_name, Field::INDEX_TOKENIZED | Field::STORE_NO); |
| doc.add(*field); |
| |
| const char* field_value_data = "人民可以得到更多实惠"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| auto* stream = analyzer->tokenStream(field->name(), stringReader.get()); |
| field->setValue(stream); |
| w.addDocument(&doc); |
| |
| const char* field_value_data1 = "中国人民银行"; |
| auto stringReader1 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data1, strlen(field_value_data1), false); |
| auto* stream1 = analyzer->tokenStream(field->name(), stringReader1.get()); |
| field->setValue(stream1); |
| w.addDocument(&doc); |
| |
| const char* field_value_data2 = "洛杉矶人，洛杉矶居民"; |
| auto stringReader2 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data2, strlen(field_value_data2), false); |
| auto* stream2 = analyzer->tokenStream(field->name(), stringReader2.get()); |
| field->setValue(stream2); |
| w.addDocument(&doc); |
| |
| const char* field_value_data3 = "民族，人民"; |
| auto stringReader3 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data3, strlen(field_value_data3), false); |
| auto* stream3 = analyzer->tokenStream(field->name(), stringReader3.get()); |
| field->setValue(stream3); |
| w.addDocument(&doc); |
| |
| w.close(); |
| doc.clear(); |
| _CLDELETE(stream) |
| _CLDELETE(stream1) |
| _CLDELETE(stream2) |
| _CLDELETE(stream3) |
| } catch (CLuceneError& r) { |
| printf("clucene error in testJiebaMatch: %s\n", r.what()); |
| } |
| IndexSearcher searcher(&dir); |
| |
| std::vector<std::string> analyse_result; |
| const char* value = "民族"; |
| auto analyzer = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(L"chinese", false); |
| auto reader = std::make_unique<lucene::util::SStringReader<char>>(value, strlen(value), false); |
| |
| lucene::analysis::TokenStream* token_stream = analyzer->tokenStream(field_name, reader.get()); |
| |
| lucene::analysis::Token token; |
| |
| while (token_stream->next(&token)) { |
| if (token.termLength<char>() != 0) { |
| analyse_result.emplace_back(token.termBuffer<char>(), token.termLength<char>()); |
| } |
| } |
| |
| if (token_stream != nullptr) { |
| token_stream->close(); |
| } |
| _CLDELETE(token_stream) |
| |
| auto query = std::make_unique<lucene::search::BooleanQuery>(); |
| for (const auto& t : analyse_result) { |
| std::wstring token_ws = StringUtil::string_to_wstring(t); |
| auto* term = _CLNEW lucene::index::Term(field_name, token_ws.c_str()); |
| dynamic_cast<lucene::search::BooleanQuery*>(query.get()) |
| ->add(_CLNEW lucene::search::TermQuery(term), true, |
| lucene::search::BooleanClause::SHOULD); |
| _CLDECDELETE(term); |
| } |
| Hits* hits1 = searcher.search(query.get()); |
| CLUCENE_ASSERT(1 == hits1->length()); |
| _CLDELETE(hits1) |
| _CLDELETE_ARRAY(field_name) |
| } |
| |
| void testJiebaMatch2(CuTest* tc) { |
| RAMDirectory dir; |
| |
| auto analyzer = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(); |
| analyzer->setLanguage(L"chinese"); |
| analyzer->setMode(lucene::analysis::AnalyzerMode::Default); |
| |
| IndexWriter w(&dir, analyzer.get(), true); |
| w.setUseCompoundFile(false); |
| auto field_name = lucene::util::Misc::_charToWide("chinese"); |
| |
| Document doc; |
| auto field = _CLNEW Field(field_name, Field::INDEX_TOKENIZED | Field::STORE_NO); |
| doc.add(*field); |
| |
| try { |
| const char* field_value_data = "人民可以得到更多实惠"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| auto* stream = analyzer->tokenStream(field->name(), stringReader.get()); |
| field->setValue(stream); |
| w.addDocument(&doc); |
| |
| const char* field_value_data1 = "中国人民银行"; |
| auto stringReader1 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data1, strlen(field_value_data1), false); |
| auto* stream1 = analyzer->tokenStream(field->name(), stringReader1.get()); |
| field->setValue(stream1); |
| w.addDocument(&doc); |
| |
| const char* field_value_data2 = "洛杉矶人，洛杉矶居民"; |
| auto stringReader2 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data2, strlen(field_value_data2), false); |
| auto* stream2 = analyzer->tokenStream(field->name(), stringReader2.get()); |
| field->setValue(stream2); |
| w.addDocument(&doc); |
| |
| const char* field_value_data3 = "民族，人民"; |
| auto stringReader3 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data3, strlen(field_value_data3), false); |
| auto* stream3 = analyzer->tokenStream(field->name(), stringReader3.get()); |
| field->setValue(stream3); |
| w.addDocument(&doc); |
| |
| w.close(); |
| doc.clear(); |
| _CLDELETE(stream) |
| _CLDELETE(stream1) |
| _CLDELETE(stream2) |
| _CLDELETE(stream3) |
| } catch (CLuceneError& r) { |
| printf("clucene error in testJiebaMatch2: %s\n", r.what()); |
| } |
| IndexSearcher searcher(&dir); |
| |
| std::vector<std::string> analyse_result; |
| const char* value = "人民"; |
| auto analyzer1 = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(L"chinese", false); |
| auto reader = std::make_unique<lucene::util::SStringReader<char>>(value, strlen(value), false); |
| |
| lucene::analysis::TokenStream* token_stream = analyzer1->tokenStream(field_name, reader.get()); |
| |
| lucene::analysis::Token token; |
| |
| while (token_stream->next(&token)) { |
| if (token.termLength<char>() != 0) { |
| analyse_result.emplace_back(token.termBuffer<char>(), token.termLength<char>()); |
| } |
| } |
| |
| if (token_stream != nullptr) { |
| token_stream->close(); |
| } |
| _CLDELETE(token_stream) |
| auto query = std::make_unique<lucene::search::BooleanQuery>(); |
| for (const auto& t : analyse_result) { |
| std::wstring token_ws = StringUtil::string_to_wstring(t); |
| auto* term = _CLNEW lucene::index::Term(field_name, token_ws.c_str()); |
| dynamic_cast<lucene::search::BooleanQuery*>(query.get()) |
| ->add(_CLNEW lucene::search::TermQuery(term), true, |
| lucene::search::BooleanClause::SHOULD); |
| _CLDECDELETE(term); |
| } |
| |
| Hits* hits1 = searcher.search(query.get()); |
| CLUCENE_ASSERT(2 == hits1->length()); |
| _CLDELETE(hits1) |
| _CLDELETE_ARRAY(field_name) |
| } |
| |
| void testJiebaMatchHuge(CuTest* tc) { |
| RAMDirectory dir; |
| |
| auto analyzer = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(); |
| analyzer->setLanguage(L"chinese"); |
| analyzer->setMode(lucene::analysis::AnalyzerMode::Default); |
| analyzer->initDict("./dict"); |
| |
| IndexWriter w(&dir, analyzer.get(), true); |
| w.setUseCompoundFile(false); |
| auto field_name = lucene::util::Misc::_charToWide("chinese"); |
| |
| Document doc; |
| auto field = _CLNEW Field(field_name, Field::INDEX_TOKENIZED | Field::STORE_NO); |
| doc.add(*field); |
| |
| const char* field_value_data = |
| "数据模型\n" |
| "本文档主要从逻辑层面，描述 Doris 的数据模型，以帮助用户更好的使用 Doris " |
| "应对不同的业务场景。\n" |
| "\n" |
| "基本概念\n" |
| "在 Doris 中，数据以表（Table）的形式进行逻辑上的描述。 " |
| "一张表包括行（Row）和列（Column）。Row 即用户的一行数据。Column " |
| "用于描述一行数据中不同的字段。\n" |
| "\n" |
| "Column 可以分为两大类：Key 和 Value。从业务角度看，Key 和 Value " |
| "可以分别对应维度列和指标列。Doris的key列是建表语句中指定的列，建表语句中的关键字\\'" |
| "unique key\\'或\\'aggregate key\\'或\\'duplicate key\\'后面的列就是 Key 列，除了 Key " |
| "列剩下的就是 Value a列。\n" |
| "\n" |
| "Doris 的数据模型主要分为3类:\n" |
| "\n" |
| "Aggregate\n" |
| "Unique\n" |
| "Duplicate\n" |
| "下面我们分别介绍。\n" |
| "\n" |
| "Aggregate 模型\n" |
| "我们以实际的例子来说明什么是聚合模型，以及如何正确的使用聚合模型。\n" |
| "\n" |
| "示例1：导入数据聚合\n" |
| "假设业务有如下数据表模式：\n" |
| "\n" |
| "ColumnName Type AggregationType Comment\n" |
| "user_id LARGEINT 用户id\n" |
| "date DATE 数据灌入日期\n" |
| "city VARCHAR(20) 用户所在城市\n" |
| "age SMALLINT 用户年龄\n" |
| "sex TINYINT 用户性别\n" |
| "last_visit_date DATETIME REPLACE 用户最后一次访问时间\n" |
| "cost BIGINT SUM 用户总消费\n" |
| "max_dwell_time INT MAX 用户最大停留时间\n" |
| "min_dwell_time INT MIN 用户最小停留时间\n" |
| "如果转换成建表语句则如下（省略建表语句中的 Partition 和 Distribution 信息）\n" |
| "\n" |
| "CREATE TABLE IF NOT EXISTS example_db.example_tbl\n" |
| "(\n" |
| "user_id LARGEINT NOT NULL COMMENT \"用户id\",\n" |
| "date DATE NOT NULL COMMENT \"数据灌入日期时间\",\n" |
| "city VARCHAR(20) COMMENT \"用户所在城市\",\n" |
| "age SMALLINT COMMENT \"用户年龄\",\n" |
| "sex TINYINT COMMENT \"用户性别\",\n" |
| "last_visit_date DATETIME REPLACE DEFAULT \"1970-01-01 00:00:00\" COMMENT " |
| "\"用户最后一次访问时间\",\n" |
| "cost BIGINT SUM DEFAULT \"0\" COMMENT \"用户总消费\",\n" |
| "max_dwell_time INT MAX DEFAULT \"0\" COMMENT \"用户最大停留时间\",\n" |
| "min_dwell_time INT MIN DEFAULT \"99999\" COMMENT \"用户最小停留时间\"\n" |
| ")\n" |
| "AGGREGATE KEY(user_id, date, city, age, sex)\n" |
| "DISTRIBUTED BY HASH(user_id) BUCKETS 1\n" |
| "PROPERTIES (\n" |
| "\"replication_allocation\" = \"tag.location.default: 1\"\n" |
| ");\n" |
| "\n" |
| "可以看到，这是一个典型的用户信息和访问行为的事实表。 " |
| "在一般星型模型中，用户信息和访问行为一般分别存放在维度表和事实表中。这里我们为了更加方" |
| "便的解释 Doris 的数据模型，将两部分信息统一存放在一张表中。\n" |
| "\n" |
| "表中的列按照是否设置了 AggregationType，分为 Key (维度列) 和 " |
| "Value（指标列）。没有设置 AggregationType 的，如 user_id、date、age ... 等称为 " |
| "Key，而设置了 AggregationType 的称为 Value。\n" |
| "\n" |
| "当我们导入数据时，对于 Key 列相同的行会聚合成一行，而 Value 列会按照设置的 " |
| "AggregationType 进行聚合。 AggregationType 目前有以下四种聚合方式：\n" |
| "\n" |
| "SUM：求和，多行的 Value 进行累加。\n" |
| "REPLACE：替代，下一批数据中的 Value 会替换之前导入过的行中的 Value。\n" |
| "MAX：保留最大值。\n" |
| "MIN：保留最小值。\n" |
| "假设我们有以下导入数据（原始数据）：\n" |
| "\n" |
| "user_id date city age sex last_visit_date cost max_dwell_time min_dwell_time\n" |
| "10000 2017-10-01 北京 20 0 2017-10-01 06:00:00 20 10 10\n" |
| "10000 2017-10-01 北京 20 0 2017-10-01 07:00:00 15 2 2\n" |
| "10001 2017-10-01 北京 30 1 2017-10-01 17:05:45 2 22 22\n" |
| "10002 2017-10-02 上海 20 1 2017-10-02 12:59:12 200 5 5\n" |
| "10003 2017-10-02 广州 32 0 2017-10-02 11:20:00 30 11 11\n" |
| "10004 2017-10-01 深圳 35 0 2017-10-01 10:00:15 100 3 3\n" |
| "10004 2017-10-03 深圳 35 0 2017-10-03 10:20:22 11 6 6\n" |
| "通过sql导入数据：\n" |
| "\n" |
| "insert into example_db.example_tbl values\n" |
| "(10000,\"2017-10-01\",\"北京\",20,0,\"2017-10-01 06:00:00\",20,10,10),\n" |
| "(10000,\"2017-10-01\",\"北京\",20,0,\"2017-10-01 07:00:00\",15,2,2),\n" |
| "(10001,\"2017-10-01\",\"北京\",30,1,\"2017-10-01 17:05:45\",2,22,22),\n" |
| "(10002,\"2017-10-02\",\"上海\",20,1,\"2017-10-02 12:59:12\",200,5,5),\n" |
| "(10003,\"2017-10-02\",\"广州\",32,0,\"2017-10-02 11:20:00\",30,11,11),\n" |
| "(10004,\"2017-10-01\",\"深圳\",35,0,\"2017-10-01 10:00:15\",100,3,3),\n" |
| "(10004,\"2017-10-03\",\"深圳\",35,0,\"2017-10-03 10:20:22\",11,6,6);\n" |
| "\n" |
| "我们假设这是一张记录用户访问某商品页面行为的表。我们以第一行数据为例，解释如下：\n" |
| "\n" |
| "数据 说明\n" |
| "10000 用户id，每个用户唯一识别id\n" |
| "2017-10-01 数据入库时间，精确到日期\n" |
| "北京 用户所在城市\n" |
| "20 用户年龄\n" |
| "0 性别男（1 代表女性）\n" |
| "2017-10-01 06:00:00 用户本次访问该页面的时间，精确到秒\n" |
| "20 用户本次访问产生的消费\n" |
| "10 用户本次访问，驻留该页面的时间\n" |
| "10 用户本次访问，驻留该页面的时间（冗余）\n" |
| "那么当这批数据正确导入到 Doris 中后，Doris 中最终存储如下：\n" |
| "\n" |
| "user_id date city age sex last_visit_date cost max_dwell_time min_dwell_time\n" |
| "10000 2017-10-01 北京 20 0 2017-10-01 07:00:00 35 10 2\n" |
| "10001 2017-10-01 北京 30 1 2017-10-01 17:05:45 2 22 22\n" |
| "10002 2017-10-02 上海 20 1 2017-10-02 12:59:12 200 5 5\n" |
| "10003 2017-10-02 广州 32 0 2017-10-02 11:20:00 30 11 11\n" |
| "10004 2017-10-01 深圳 35 0 2017-10-01 10:00:15 100 3 3\n" |
| "10004 2017-10-03 深圳 35 0 2017-10-03 10:20:22 11 6 6\n" |
| "可以看到，用户 10000 " |
| "只剩下了一行聚合后的数据。而其余用户的数据和原始数据保持一致。这里先解释下用户 10000 " |
| "聚合后的数据：\n" |
| "\n" |
| "前5列没有变化，从第6列 last_visit_date 开始：\n" |
| "\n" |
| "2017-10-01 07:00:00：因为 last_visit_date 列的聚合方式为 REPLACE，所以 2017-10-01 " |
| "07:00:00 替换了 2017-10-01 06:00:00 保存了下来。\n" |
| "\n" |
| "注：在同一个导入批次中的数据，对于 REPLACE " |
| "这种聚合方式，替换顺序不做保证。如在这个例子中，最终保存下来的，也有可能是 2017-10-01 " |
| "06:00:00。而对于不同导入批次中的数据，可以保证，后一批次的数据会替换前一批次。\n" |
| "\n" |
| "35：因为 cost 列的聚合类型为 SUM，所以由 20 + 15 累加获得 35。\n" |
| "\n" |
| "10：因为 max_dwell_time 列的聚合类型为 MAX，所以 10 和 2 取最大值，获得 10。\n" |
| "\n" |
| "2：因为 min_dwell_time 列的聚合类型为 MIN，所以 10 和 2 取最小值，获得 2。\n" |
| "\n" |
| "经过聚合，Doris " |
| "中最终只会存储聚合后的数据。换句话说，即明细数据会丢失，用户不能够再查询到聚合前的明细" |
| "数据了。\n" |
| "\n" |
| "示例2：保留明细数据\n" |
| "接示例1，我们将表结构修改如下：\n" |
| "\n" |
| "ColumnName Type AggregationType Comment\n" |
| "user_id LARGEINT 用户id\n" |
| "date DATE 数据灌入日期\n" |
| "timestamp DATETIME 数据灌入时间，精确到秒\n" |
| "city VARCHAR(20) 用户所在城市\n" |
| "age SMALLINT 用户年龄\n" |
| "sex TINYINT 用户性别\n" |
| "last_visit_date DATETIME REPLACE 用户最后一次访问时间\n" |
| "cost BIGINT SUM 用户总消费\n" |
| "max_dwell_time INT MAX 用户最大停留时间\n" |
| "min_dwell_time INT MIN 用户最小停留时间\n" |
| "即增加了一列 timestamp，记录精确到秒的数据灌入时间。 同时，将AGGREGATE " |
| "KEY设置为AGGREGATE KEY(user_id, date, timestamp, city, age, sex)\n" |
| "\n" |
| "导入数据如下：\n" |
| "\n" |
| "user_id date timestamp city age sex last_visit_date cost max_dwell_time " |
| "min_dwell_time\n" |
| "10000 2017-10-01 2017-10-01 08:00:05 北京 20 0 2017-10-01 06:00:00 20 10 10\n" |
| "10000 2017-10-01 2017-10-01 09:00:05 北京 20 0 2017-10-01 07:00:00 15 2 2\n" |
| "10001 2017-10-01 2017-10-01 18:12:10 北京 30 1 2017-10-01 17:05:45 2 22 22\n" |
| "10002 2017-10-02 2017-10-02 13:10:00 上海 20 1 2017-10-02 12:59:12 200 5 5\n" |
| "10003 2017-10-02 2017-10-02 13:15:00 广州 32 0 2017-10-02 11:20:00 30 11 11\n" |
| "10004 2017-10-01 2017-10-01 12:12:48 深圳 35 0 2017-10-01 10:00:15 100 3 3\n" |
| "10004 2017-10-03 2017-10-03 12:38:20 深圳 35 0 2017-10-03 10:20:22 11 6 6\n" |
| "通过sql导入数据：\n" |
| "\n" |
| "insert into example_db.example_tbl values\n" |
| "(10000,\"2017-10-01\",\"2017-10-01 08:00:05\",\"北京\",20,0,\"2017-10-01 " |
| "06:00:00\",20,10,10),\n" |
| "(10000,\"2017-10-01\",\"2017-10-01 09:00:05\",\"北京\",20,0,\"2017-10-01 " |
| "07:00:00\",15,2,2),\n" |
| "(10001,\"2017-10-01\",\"2017-10-01 18:12:10\",\"北京\",30,1,\"2017-10-01 " |
| "17:05:45\",2,22,22),\n" |
| "(10002,\"2017-10-02\",\"2017-10-02 13:10:00\",\"上海\",20,1,\"2017-10-02 " |
| "12:59:12\",200,5,5),\n" |
| "(10003,\"2017-10-02\",\"2017-10-02 13:15:00\",\"广州\",32,0,\"2017-10-02 " |
| "11:20:00\",30,11,11),\n" |
| "(10004,\"2017-10-01\",\"2017-10-01 12:12:48\",\"深圳\",35,0,\"2017-10-01 " |
| "10:00:15\",100,3,3),\n" |
| "(10004,\"2017-10-03\",\"2017-10-03 12:38:20\",\"深圳\",35,0,\"2017-10-03 " |
| "10:20:22\",11,6,6);\n" |
| "\n" |
| "那么当这批数据正确导入到 Doris 中后，Doris 中最终存储如下：\n" |
| "\n" |
| "user_id date timestamp city age sex last_visit_date cost max_dwell_time " |
| "min_dwell_time\n" |
| "10000 2017-10-01 2017-10-01 08:00:05 北京 20 0 2017-10-01 06:00:00 20 10 10\n" |
| "10000 2017-10-01 2017-10-01 09:00:05 北京 20 0 2017-10-01 07:00:00 15 2 2\n" |
| "10001 2017-10-01 2017-10-01 18:12:10 北京 30 1 2017-10-01 17:05:45 2 22 22\n" |
| "10002 2017-10-02 2017-10-02 13:10:00 上海 20 1 2017-10-02 12:59:12 200 5 5\n" |
| "10003 2017-10-02 2017-10-02 13:15:00 广州 32 0 2017-10-02 11:20:00 30 11 11\n" |
| "10004 2017-10-01 2017-10-01 12:12:48 深圳 35 0 2017-10-01 10:00:15 100 3 3\n" |
| "10004 2017-10-03 2017-10-03 12:38:20 深圳 35 0 2017-10-03 10:20:22 11 6 6\n" |
| "我们可以看到，存储的数据，和导入数据完全一样，没有发生任何聚合。这是因为，这批数据中，" |
| "因为加入了 timestamp 列，所有行的 Key " |
| "都不完全相同。也就是说，只要保证导入的数据中，每一行的 Key " |
| "都不完全相同，那么即使在聚合模型下，Doris 也可以保存完整的明细数据。\n" |
| "\n" |
| "示例3：导入数据与已有数据聚合\n" |
| "接示例1。假设现在表中已有数据如下：\n" |
| "\n" |
| "user_id date city age sex last_visit_date cost max_dwell_time min_dwell_time\n" |
| "10000 2017-10-01 北京 20 0 2017-10-01 07:00:00 35 10 2\n" |
| "10001 2017-10-01 北京 30 1 2017-10-01 17:05:45 2 22 22\n" |
| "10002 2017-10-02 上海 20 1 2017-10-02 12:59:12 200 5 5\n" |
| "10003 2017-10-02 广州 32 0 2017-10-02 11:20:00 30 11 11\n" |
| "10004 2017-10-01 深圳 35 0 2017-10-01 10:00:15 100 3 3\n" |
| "10004 2017-10-03 深圳 35 0 2017-10-03 10:20:22 11 6 6\n" |
| "我们再导入一批新的数据：\n" |
| "\n" |
| "user_id date city age sex last_visit_date cost max_dwell_time min_dwell_time\n" |
| "10004 2017-10-03 深圳 35 0 2017-10-03 11:22:00 44 19 19\n" |
| "10005 2017-10-03 长沙 29 1 2017-10-03 18:11:02 3 1 1\n" |
| "通过sql导入数据：\n" |
| "\n" |
| "insert into example_db.example_tbl values\n" |
| "(10004,\"2017-10-03\",\"深圳\",35,0,\"2017-10-03 11:22:00\",44,19,19),\n" |
| "(10005,\"2017-10-03\",\"长沙\",29,1,\"2017-10-03 18:11:02\",3,1,1);\n" |
| "\n" |
| "那么当这批数据正确导入到 Doris 中后，Doris 中最终存储如下：\n" |
| "\n" |
| "user_id date city age sex last_visit_date cost max_dwell_time min_dwell_time\n" |
| "10000 2017-10-01 北京 20 0 2017-10-01 07:00:00 35 10 2\n" |
| "10001 2017-10-01 北京 30 1 2017-10-01 17:05:45 2 22 22\n" |
| "10002 2017-10-02 上海 20 1 2017-10-02 12:59:12 200 5 5\n" |
| "10003 2017-10-02 广州 32 0 2017-10-02 11:20:00 30 11 11\n" |
| "10004 2017-10-01 深圳 35 0 2017-10-01 10:00:15 100 3 3\n" |
| "10004 2017-10-03 深圳 35 0 2017-10-03 11:22:00 55 19 6\n" |
| "10005 2017-10-03 长沙 29 1 2017-10-03 18:11:02 3 1 1\n" |
| "可以看到，用户 10004 的已有数据和新导入的数据发生了聚合。同时新增了 10005 " |
| "用户的数据。\n" |
| "\n" |
| "数据的聚合，在 Doris 中有如下三个阶段发生：\n" |
| "\n" |
| "每一批次数据导入的 ETL 阶段。该阶段会在每一批次导入的数据内部进行聚合。\n" |
| "底层 BE 进行数据 Compaction 的阶段。该阶段，BE " |
| "会对已导入的不同批次的数据进行进一步的聚合。\n" |
| "数据查询阶段。在数据查询时，对于查询涉及到的数据，会进行对应的聚合。\n" |
| "数据在不同时间，可能聚合的程度不一致。比如一批数据刚导入时，可能还未与之前已存在的数据" |
| "进行聚合。但是对于用户而言，用户只能查询到聚合后的数据。即不同的聚合程度对于用户查询而" |
| "言是透明的。用户需始终认为数据以最终的完成的聚合程度存在，而不应假设某些聚合还未发生。" |
| "（可参阅聚合模型的局限性一节获得更多详情。）\n" |
| "\n" |
| "Unique 模型\n" |
| "在某些多维分析场景下，用户更关注的是如何保证 Key 的唯一性，即如何获得 Primary Key " |
| "唯一性约束。因此，我们引入了 /；·90Unique " |
| "数据模型。在1." |
| "2版本之前，该模型本质上是聚合模型的一个特例，也是一种简化的表结构表示方式。由于聚合模" |
| "型的实现方式是读时合并（merge on " |
| "read)，因此在一些聚合查询上性能不佳（参考后续章节聚合模型的局限性的描述），在1." |
| "2版本我们引入了Unique模型新的实现方式，写时合并（merge on " |
| "write），通过在写入时做一些额外的工作，实现了最优的查询性能。写时合并将在未来替换读时" |
| "合并成为Unique模型的默认实现方式，两者将会短暂的共存一段时间。下面将对两种实现方式分别" |
| "举例进行说明。\n" |
| "\n" |
| "读时合并（与聚合模型相同的实现方式）\n" |
| "ColumnName Type IsKey Comment\n" |
| "user_id BIGINT Yes 用户id\n" |
| "username VARCHAR(50) Yes 用户昵称\n" |
| "city VARCHAR(20) No 用户所在城市\n" |
| "age SMALLINT No 用户年龄\n" |
| "sex TINYINT No 用户性别\n" |
| "phone LARGEINT No 用户电话\n" |
| "address VARCHAR(500) No 用户住址\n" |
| "register_time DATETIME No 用户注册时间\n" |
| "这是一个典型的用户基础信息表。这类数据没有聚合需求，只需保证主键唯一性。（这里的主键为" |
| " user_id + username）。那么我们的建表语句如下：\n" |
| "\n" |
| "CREATE TABLE IF NOT EXISTS example_db.example_tbl\n" |
| "(\n" |
| "user_id LARGEINT NOT NULL COMMENT \"用户id\",\n" |
| "username VARCHAR(50) NOT NULL COMMENT \"用户昵称\",\n" |
| "city VARCHAR(20) COMMENT \"用户所在城市\",\n" |
| "age SMALLINT COMMENT \"用户年龄\",\n" |
| "sex TINYINT COMMENT \"用户性别\",\n" |
| "phone LARGEINT COMMENT \"用户电话\",\n" |
| "address VARCHAR(500) COMMENT \"用户地址\",\n" |
| "register_time DATETIME COMMENT \"用户注册时间\"\n" |
| ")\n" |
| "UNIQUE KEY(user_id, username)\n" |
| "DISTRIBUTED BY HASH(user_id) BUCKETS 1\n" |
| "PROPERTIES (\n" |
| "\"replication_allocation\" = \"tag.location.default: 1\"\n" |
| ");\n" |
| "\n" |
| "而这个表结构，完全同等于以下使用聚合模型描述的表结构：\n" |
| "\n" |
| "ColumnName Type AggregationType Comment\n" |
| "user_id BIGINT 用户id\n" |
| "username VARCHAR(50) 用户昵称\n" |
| "city VARCHAR(20) REPLACE 用户所在城市\n" |
| "age SMALLINT REPLACE 用户年龄\n" |
| "sex TINYINT REPLACE 用户性别\n" |
| "phone LARGEINT REPLACE 用户电话\n" |
| "address VARCHAR(500) REPLACE 用户住址\n" |
| "register_time DATETIME REPLACE 用户注册时间\n" |
| "及建表语句：\n" |
| "\n" |
| "CREATE TABLE IF NOT EXISTS example_db.example_tbl\n" |
| "(\n" |
| "user_id LARGEINT NOT NULL COMMENT \"用户id\",\n" |
| "username VARCHAR(50) NOT NULL COMMENT \"用户昵称\",\n" |
| "city VARCHAR(20) REPLACE COMMENT \"用户所在城市\",\n" |
| "age SMALLINT REPLACE COMMENT \"用户年龄\",\n" |
| "sex TINYINT REPLACE COMMENT \"用户性别\",\n" |
| "phone LARGEINT REPLACE COMMENT \"用户电话\",\n" |
| "address VARCHAR(500) REPLACE COMMENT \"用户地址\",\n" |
| "register_time DATETIME REPLACE COMMENT \"用户注册时间\"\n" |
| ")\n" |
| "AGGREGATE KEY(user_id, username)\n" |
| "DISTRIBUTED BY HASH(user_id) BUCKETS 1\n" |
| "PROPERTIES (\n" |
| "\"replication_allocation\" = \"tag.location.default: 1\"\n" |
| ");\n" |
| "\n" |
| "即Unique 模型的读时合并实现完全可以用聚合模型中的 REPLACE " |
| "方式替代。其内部的实现方式和数据存储方式也完全一样。这里不再继续举例说明。\n" |
| "\n" |
| "SinceVersion 1.2\n" |
| "写时合并\n" |
| "Unqiue模型的写时合并实现，与聚合模型就是完全不同的两种模型了，查询性能更接近于duplicat" |
| "e模型，在有主键约束需求的场景上相比聚合模型有较大的查询性能优势，尤其是在聚合查询以及" |
| "需要用索引过滤大量数据的查询中。\n" |
| "\n" |
| "在 1.2.0 " |
| "版本中，作为一个新的feature，写时合并默认关闭，用户可以通过添加下面的property来开启\n" |
| "\n" |
| "\"enable_unique_key_merge_on_write\" = \"true\"\n" |
| "\n" |
| "仍然以上面的表为例，建表语句为\n" |
| "\n" |
| "CREATE TABLE IF NOT EXISTS example_db.example_tbl\n" |
| "(\n" |
| "user_id LARGEINT NOT NULL COMMENT \"用户id\",\n" |
| "username VARCHAR(50) NOT NULL COMMENT \"用户昵称\",\n" |
| "city VARCHAR(20) COMMENT \"用户所在城市\",\n" |
| "age SMALLINT COMMENT \"用户年龄\",\n" |
| "sex TINYINT COMMENT \"用户性别\",\n" |
| "phone LARGEINT COMMENT \"用户电话\",\n" |
| "address VARCHAR(500) COMMENT \"用户地址\",\n" |
| "register_time DATETIME COMMENT \"用户注册时间\"\n" |
| ")\n" |
| "UNIQUE KEY(user_id, username)\n" |
| "DISTRIBUTED BY HASH(user_id) BUCKETS 1\n" |
| "PROPERTIES (\n" |
| "\"replication_allocation\" = \"tag.location.default: 1\",\n" |
| "\"enable_unique_key_merge_on_write\" = \"true\"\n" |
| ");\n" |
| "\n" |
| "使用这种建表语句建出来的表结构，与聚合模型就完全不同了：\n" |
| "\n" |
| "ColumnName Type AggregationType Comment\n" |
| "user_id BIGINT 用户id\n" |
| "username VARCHAR(50) 用户昵称\n" |
| "city VARCHAR(20) NONE 用户所在城市\n" |
| "age SMALLINT NONE 用户年龄\n" |
| "sex TINYINT NONE 用户性别\n" |
| "phone LARGEINT NONE 用户电话\n" |
| "address VARCHAR(500) NONE 用户住址\n" |
| "register_time DATETIME NONE 用户注册时间\n" |
| "在开启了写时合并选项的Unique表上，数据在导入阶段就会去将被覆盖和被更新的数据进行标记删" |
| "除，同时将新的数据写入新的文件。在查询的时候，所有被标记删除的数据都会在文件级别被过滤" |
| "掉，读取出来的数据就都是最新的数据，消除掉了读时合并中的数据聚合过程，并且能够在很多情" |
| "况下支持多种谓词的下推。因此在许多场景都能带来比较大的性能提升，尤其是在有聚合查询的情" |
| "况下。\n" |
| "\n" |
| "【注意】\n" |
| "\n" |
| "新的Merge-on-write实现默认关闭，且只能在建表时通过指定property的方式打开。\n" |
| "旧的Merge-on-" |
| "read的实现无法无缝升级到新版本的实现（数据组织方式完全不同），如果需要改为使用写时合并" |
| "的实现版本，需要手动执行insert into unique-mow-table select * from source table.\n" |
| "在Unique模型上独有的delete sign 和 sequence " |
| "col，在写时合并的新版实现中仍可以正常使用，用法没有变化。\n" |
| "Duplicate 模型\n" |
| "在某些多维分析场景下，数据既没有主键，也没有聚合需求。因此，我们引入 Duplicate " |
| "数据模型来满足这类需求。举例说明。\n" |
| "\n" |
| "ColumnName Type SortKey Comment\n" |
| "timestamp DATETIME Yes 日志时间\n" |
| "type INT Yes 日志类型\n" |
| "error_code INT Yes 错误码\n" |
| "error_msg VARCHAR(1024) No 错误详细信息\n" |
| "op_id BIGINT No 负责人id\n" |
| "op_time DATETIME No 处理时间\n" |
| "建表语句如下：\n" |
| "\n" |
| "CREATE TABLE IF NOT EXISTS example_db.example_tbl\n" |
| "(\n" |
| "timestamp DATETIME NOT NULL COMMENT \"日志时间\",\n" |
| "type INT NOT NULL COMMENT \"日志类型\",\n" |
| "error_code INT COMMENT \"错误码\",\n" |
| "error_msg VARCHAR(1024) COMMENT \"错误详细信息\",\n" |
| "op_id BIGINT COMMENT \"负责人id\",\n" |
| "op_time DATETIME COMMENT \"处理时间\"\n" |
| ")\n" |
| "DUPLICATE KEY(timestamp, type, error_code)\n" |
| "DISTRIBUTED BY HASH(type) BUCKETS 1\n" |
| "PROPERTIES (\n" |
| "\"replication_allocation\" = \"tag.location.default: 1\"\n" |
| ");\n" |
| "\n" |
| "这种数据模型区别于 Aggregate 和 Unique " |
| "模型。数据完全按照导入文件中的数据进行存储，不会有任何聚合。即使两行数据完全相同，也都" |
| "会保留。 而在建表语句中指定的 DUPLICATE " |
| "KEY，只是用来指明底层数据按照那些列进行排序。（更贴切的名称应该为 “Sorted " |
| "Column”，这里取名 “DUPLICATE KEY” 只是用以明确表示所用的数据模型。关于 “Sorted " |
| "Column”的更多解释，可以参阅前缀索引）。在 DUPLICATE KEY " |
| "的选择上，我们建议适当的选择前 2-4 列就可以。\n" |
| "\n" |
| "这种数据模型适用于既没有聚合需求，又没有主键唯一性约束的原始数据的存储。更多使用场景，" |
| "可参阅聚合模型的局限性小节。\n" |
| "\n" |
| "聚合模型的局限性\n" |
| "这里我们针对 Aggregate 模型，来介绍下聚合模型的局限性。\n" |
| "\n" |
| "在聚合模型中，模型对外展现的，是最终聚合后的数据。也就是说，任何还未聚合的数据（比如说" |
| "两个不同导入批次的数据），必须通过某种方式，以保证对外展示的一致性。我们举例说明。\n" |
| "\n" |
| "假设表结构如下：\n" |
| "\n" |
| "ColumnName Type AggregationType Comment\n" |
| "user_id LARGEINT 用户id\n" |
| "date DATE 数据灌入日期\n" |
| "cost BIGINT SUM 用户总消费\n" |
| "假设存储引擎中有如下两个已经导入完成的批次的数据：\n" |
| "\n" |
| "batch 1\n" |
| "\n" |
| "user_id date cost\n" |
| "10001 2017-11-20 50\n" |
| "10002 2017-11-21 39\n" |
| "batch 2\n" |
| "\n" |
| "user_id date cost\n" |
| "10001 2017-11-20 1\n" |
| "10001 2017-11-21 5\n" |
| "10003 2017-11-22 22\n" |
| "可以看到，用户 10001 " |
| "分属在两个导入批次中的数据还没有聚合。但是为了保证用户只能查询到如下最终聚合后的数据：" |
| "\n" |
| "\n" |
| "user_id date cost\n" |
| "10001 2017-11-20 51\n" |
| "10001 2017-11-21 5\n" |
| "10002 2017-11-21 39\n" |
| "10003 2017-11-22 22\n" |
| "我们在查询引擎中加入了聚合算子，来保证数据对外的一致性。\n" |
| "\n" |
| "另外，在聚合列（Value）上，执行与聚合类型不一致的聚合类查询时，要注意语意。比如我们在" |
| "如上示例中执行如下查询：\n" |
| "\n" |
| "SELECT MIN(cost) FROM table;\n" |
| "\n" |
| "得到的结果是 5，而不是 1。\n" |
| "\n" |
| "同时，这种一致性保证，在某些查询中，会极大的降低查询效率。\n" |
| "\n" |
| "我们以最基本的 count(*) 查询为例：\n" |
| "\n" |
| "SELECT COUNT(*) FROM table;\n" |
| "\n" |
| "在其他数据库中，这类查询都会很快的返回结果。因为在实现上，我们可以通过如“导入时对行进" |
| "行计数，保存 count 的统计信息”，或者在查询时“仅扫描某一列数据，获得 count " |
| "值”的方式，只需很小的开销，即可获得查询结果。但是在 Doris " |
| "的聚合模型中，这种查询的开销非常大。\n" |
| "\n" |
| "我们以刚才的数据为例：\n" |
| "\n" |
| "batch 1\n" |
| "\n" |
| "user_id date cost\n" |
| "10001 2017-11-20 50\n" |
| "10002 2017-11-21 39\n" |
| "batch 2\n" |
| "\n" |
| "user_id date cost\n" |
| "10001 2017-11-20 1\n" |
| "10001 2017-11-21 5\n" |
| "10003 2017-11-22 22\n" |
| "因为最终的聚合结果为：\n" |
| "\n" |
| "user_id date cost\n" |
| "10001 2017-11-20 51\n" |
| "10001 2017-11-21 5\n" |
| "10002 2017-11-21 39\n" |
| "10003 2017-11-22 22\n" |
| "所以，select count(*) from table; 的正确结果应该为 4。但如果我们只扫描 user_id " |
| "这一列，如果加上查询时聚合，最终得到的结果是 3（10001, 10002, " |
| "10003）。而如果不加查询时聚合，则得到的结果是 " |
| "5（两批次一共5行数据）。可见这两个结果都是不对的。\n" |
| "\n" |
| "为了得到正确的结果，我们必须同时读取 user_id 和 date " |
| "这两列的数据，再加上查询时聚合，才能返回 4 这个正确的结果。也就是说，在 count() " |
| "查询中，Doris 必须扫描所有的 AGGREGATE KEY 列（这里就是 user_id 和 " |
| "date），并且聚合后，才能得到语意正确的结果。当聚合列非常多时，count() " |
| "查询需要扫描大量的数据。\n" |
| "\n" |
| "因此，当业务上有频繁的 count() 查询时，我们建议用户通过增加一个值恒为 1 " |
| "的，聚合类型为 SUM 的列来模拟 count()。如刚才的例子中的表结构，我们修改如下：\n" |
| "\n" |
| "ColumnName Type AggregateType Comment\n" |
| "user_id BIGINT 用户id\n" |
| "date DATE 数据灌入日期\n" |
| "cost BIGINT SUM 用户总消费\n" |
| "count BIGINT SUM 用于计算count\n" |
| "增加一个 count 列，并且导入数据中，该列值恒为 1。则 select count() from table; " |
| "的结果等价于 select sum(count) from " |
| "table;" |
| "。而后者的查询效率将远高于前者。不过这种方式也有使用限制，就是用户需要自行保证，不会重" |
| "复导入 AGGREGATE KEY 列都相同的行。否则，select sum(count) from table; " |
| "只能表述原始导入的行数，而不是 select count() from table; 的语义。\n" |
| "\n" |
| "另一种方式，就是 将如上的 count 列的聚合类型改为 REPLACE，且依然值恒为 1。那么 select " |
| "sum(count) from table; 和 select count(*) from table; " |
| "的结果将是一致的。并且这种方式，没有导入重复行的限制。\n" |
| "\n" |
| "Unique模型的写时合并实现\n" |
| "Unique模型的写时合并实现没有聚合模型的局限性，还是以刚才的数据为例，写时合并为每次导入" |
| "的rowset增加了对应的delete bitmap，来标记哪些数据被覆盖。第一批数据导入后状态如下\n" |
| "\n" |
| "batch 1\n" |
| "\n" |
| "user_id date cost delete bit\n" |
| "10001 2017-11-20 50 false\n" |
| "10002 2017-11-21 39 false\n" |
| "当第二批数据导入完成后，第一批数据中重复的行就会被标记为已删除，此时两批数据状态如下\n" |
| "\n" |
| "batch 1\n" |
| "\n" |
| "user_id date cost delete bit\n" |
| "10001 2017-11-20 50 true\n" |
| "10002 2017-11-21 39 false\n" |
| "batch 2\n" |
| "\n" |
| "user_id date cost delete bit\n" |
| "10001 2017-11-20 1 false\n" |
| "10001 2017-11-21 5 false\n" |
| "10003 2017-11-22 22 false\n" |
| "在查询时，所有在delete " |
| "bitmap中被标记删除的数据都不会读出来，因此也无需进行做任何数据聚合，上述数据中有效的行" |
| "数为4行，查询出的结果也应该是4行，也就可以采取开销最小的方式来获取结果，即前面提到的“" |
| "仅扫描某一列数据，获得 count 值”的方式。\n" |
| "\n" |
| "在测试环境中，count(*) " |
| "查询在Unique模型的写时合并实现上的性能，相比聚合模型有10倍以上的提升。\n" |
| "\n" |
| "Duplicate 模型\n" |
| "Duplicate 模型没有聚合模型的这个局限性。因为该模型不涉及聚合语意，在做 count(*) " |
| "查询时，任意选择一列查询，即可得到语意正确的结果。\n" |
| "\n" |
| "key 列\n" |
| "Duplicate、Aggregate、Unique 模型，都会在建表指定 key " |
| "列，然而实际上是有所区别的：对于 Duplicate 模型，表的key列，可以认为只是 " |
| "“排序列”，并非起到唯一标识的作用。而 Aggregate、Unique 模型这种聚合类型的表，key " |
| "列是兼顾 “排序列” 和 “唯一标识列”，是真正意义上的“ key 列”。\n" |
| "\n" |
| "数据模型的选择建议\n" |
| "因为数据模型在建表时就已经确定，且无法修改。所以，选择一个合适的数据模型非常重要。\n" |
| "\n" |
| "Aggregate " |
| "模型可以通过预聚合，极大地降低聚合查询时所需扫描的数据量和查询的计算量，非常适合有固定" |
| "模式的报表类查询场景。但是该模型对 count(*) 查询很不友好。同时因为固定了 Value " |
| "列上的聚合方式，在进行其他类型的聚合查询时，需要考虑语意正确性。\n" |
| "Unique 模型针对需要唯一主键约束的场景，可以保证主键唯一性约束。但是无法利用 ROLLUP " |
| "等预聚合带来的查询优势。\n" |
| "对于聚合查询有较高性能需求的用户，推荐使用自1.2版本加入的写时合并实现。\n" |
| "Unique " |
| "模型仅支持整行更新，如果用户既需要唯一主键约束，又需要更新部分列（例如将多张源表导入到" |
| "一张 doris 表的情形），则可以考虑使用 Aggregate 模型，同时将非主键列的聚合类型设置为 " |
| "REPLACE_IF_NOT_NULL。具体的用法可以参考语法手册\n" |
| "Duplicate 适合任意维度的 Ad-hoc " |
| "查询。虽然同样无法利用预聚合的特性，但是不受聚合模型的约束，可以发挥列存模型的优势（只" |
| "读取相关列，而不需要读取所有 Key 列）。"; |
| try { |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| auto* stream = analyzer->tokenStream(field->name(), stringReader.get()); |
| field->setValue(stream); |
| w.addDocument(&doc); |
| |
| w.close(); |
| doc.clear(); |
| _CLDELETE(stream) |
| } catch (CLuceneError& r) { |
| printf("clucene error in testJiebaMatchHuge: %s\n", r.what()); |
| } |
| |
| IndexSearcher searcher(&dir); |
| |
| std::vector<std::string> analyse_result; |
| const char* value = "相关"; |
| auto analyzer1 = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(L"chinese", false); |
| auto reader = std::make_unique<lucene::util::SStringReader<char>>(value, strlen(value), false); |
| |
| lucene::analysis::TokenStream* token_stream = analyzer1->tokenStream(field_name, reader.get()); |
| |
| lucene::analysis::Token token; |
| |
| while (token_stream->next(&token)) { |
| if (token.termLength<char>() != 0) { |
| analyse_result.emplace_back(token.termBuffer<char>(), token.termLength<char>()); |
| } |
| } |
| |
| if (token_stream != nullptr) { |
| token_stream->close(); |
| } |
| _CLDELETE(token_stream) |
| auto query = std::make_unique<lucene::search::BooleanQuery>(); |
| for (const auto& t : analyse_result) { |
| std::wstring token_ws = StringUtil::string_to_wstring(t); |
| auto* term = _CLNEW lucene::index::Term(field_name, token_ws.c_str()); |
| dynamic_cast<lucene::search::BooleanQuery*>(query.get()) |
| ->add(_CLNEW lucene::search::TermQuery(term), true, |
| lucene::search::BooleanClause::SHOULD); |
| _CLDECDELETE(term); |
| } |
| |
| Hits* hits1 = searcher.search(query.get()); |
| CLUCENE_ASSERT(1 == hits1->length()); |
| _CLDELETE(hits1) |
| _CLDELETE_ARRAY(field_name) |
| } |
| |
| void testChineseMatch(CuTest* tc) { |
| RAMDirectory dir; |
| auto* field_name = lucene::util::Misc::_charToWide("chinese"); |
| auto analyzer = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(); |
| analyzer->setLanguage(L"cjk"); |
| try { |
| IndexWriter w(&dir, analyzer.get(), true); |
| w.setUseCompoundFile(false); |
| |
| Document doc; |
| auto field = _CLNEW Field(field_name, Field::INDEX_TOKENIZED | Field::STORE_NO); |
| doc.add(*field); |
| |
| const char* field_value_data = "人民可以得到更多实惠"; |
| auto stringReader = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data, strlen(field_value_data), false); |
| auto* stream = analyzer->tokenStream(field->name(), stringReader.get()); |
| field->setValue(stream); |
| w.addDocument(&doc); |
| |
| const char* field_value_data1 = "中国人民银行"; |
| auto stringReader1 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data1, strlen(field_value_data1), false); |
| auto* stream1 = analyzer->tokenStream(field->name(), stringReader1.get()); |
| field->setValue(stream1); |
| w.addDocument(&doc); |
| |
| const char* field_value_data2 = "洛杉矶人，洛杉矶居民"; |
| auto stringReader2 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data2, strlen(field_value_data2), false); |
| auto* stream2 = analyzer->tokenStream(field->name(), stringReader2.get()); |
| field->setValue(stream2); |
| w.addDocument(&doc); |
| |
| const char* field_value_data3 = "民族，人民"; |
| auto stringReader3 = std::make_unique<lucene::util::SStringReader<char>>( |
| field_value_data3, strlen(field_value_data3), false); |
| auto* stream3 = analyzer->tokenStream(field->name(), stringReader3.get()); |
| field->setValue(stream3); |
| w.addDocument(&doc); |
| |
| w.close(); |
| doc.clear(); |
| _CLDELETE(stream) |
| _CLDELETE(stream1) |
| _CLDELETE(stream2) |
| _CLDELETE(stream3) |
| } catch (const CLuceneError& e) { |
| std::cout << "clucene error in testChineseMatch:" << e.what(); |
| } |
| IndexSearcher searcher(&dir); |
| |
| std::vector<std::string> analyse_result; |
| const char* value = "民族"; |
| auto analyzer1 = std::make_unique<lucene::analysis::LanguageBasedAnalyzer>(L"cjk", false); |
| auto reader = std::make_unique<lucene::util::SStringReader<char>>(value, strlen(value), false); |
| lucene::analysis::TokenStream* token_stream = analyzer1->tokenStream(field_name, reader.get()); |
| |
| lucene::analysis::Token token; |
| |
| while (token_stream->next(&token)) { |
| if (token.termLength<char>() != 0) { |
| analyse_result.emplace_back(token.termBuffer<char>(), token.termLength<char>()); |
| } |
| } |
| |
| if (token_stream != nullptr) { |
| token_stream->close(); |
| } |
| _CLDELETE(token_stream) |
| auto query = std::make_unique<lucene::search::BooleanQuery>(); |
| for (const auto& t : analyse_result) { |
| std::wstring token_ws = StringUtil::string_to_wstring(t); |
| auto* term = _CLNEW lucene::index::Term(field_name, token_ws.c_str()); |
| dynamic_cast<lucene::search::BooleanQuery*>(query.get()) |
| ->add(_CLNEW lucene::search::TermQuery(term), true, |
| lucene::search::BooleanClause::SHOULD); |
| _CLDECDELETE(term); |
| } |
| |
| Hits* hits1 = searcher.search(query.get()); |
| CLUCENE_ASSERT(1 == hits1->length()); |
| _CLDELETE(hits1) |
| _CLDELETE_ARRAY(field_name) |
| } |
| |
| void testLanguageBasedAnalyzer(CuTest* tc) { |
| LanguageBasedAnalyzer a; |
| CL_NS(util)::StringReader reader(_T("he abhorred accentueren")); |
| reader.mark(50); |
| TokenStream* ts; |
| Token t; |
| |
| //test with english |
| a.setLanguage(_T("English")); |
| a.setStem(false); |
| ts = a.tokenStream(_T("contents"), &reader); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(_tcscmp(t.termBuffer<TCHAR>(), _T("he")) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(_tcscmp(t.termBuffer<TCHAR>(), _T("abhorred")) == 0); |
| _CLDELETE(ts); |
| |
| //now test with dutch |
| reader.reset(0); |
| a.setLanguage(_T("Dutch")); |
| a.setStem(true); |
| ts = a.tokenStream(_T("contents"), &reader); |
| |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(_tcscmp(t.termBuffer<TCHAR>(), _T("he")) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(_tcscmp(t.termBuffer<TCHAR>(), _T("abhorred")) == 0); |
| CLUCENE_ASSERT(ts->next(&t) != NULL); |
| CLUCENE_ASSERT(_tcscmp(t.termBuffer<TCHAR>(), _T("accentuer")) == 0); |
| _CLDELETE(ts); |
| } |
| |
| CuSuite* testchinese(void) { |
| CuSuite* suite = CuSuiteNew(_T("CLucene chinese tokenizer Test")); |
| |
| SUITE_ADD_TEST(suite, testFile); |
| SUITE_ADD_TEST(suite, testCJK); |
| SUITE_ADD_TEST(suite, testLanguageBasedAnalyzer); |
| SUITE_ADD_TEST(suite, testChineseAnalyzer); |
| SUITE_ADD_TEST(suite, testSimpleJiebaTokenizer); |
| SUITE_ADD_TEST(suite, testSimpleJiebaTokenizer2); |
| SUITE_ADD_TEST(suite, testSimpleJiebaTokenizer3); |
| SUITE_ADD_TEST(suite, testSimpleJiebaTokenizer4); |
| SUITE_ADD_TEST(suite, testChineseMatch); |
| SUITE_ADD_TEST(suite, testJiebaMatch); |
| SUITE_ADD_TEST(suite, testJiebaMatch2); |
| SUITE_ADD_TEST(suite, testJiebaMatchHuge); |
| SUITE_ADD_TEST(suite, testSimpleJiebaAllModeTokenizer); |
| SUITE_ADD_TEST(suite, testSimpleJiebaDefaultModeTokenizer); |
| SUITE_ADD_TEST(suite, testSimpleJiebaDefaultModeTokenizer2); |
| SUITE_ADD_TEST(suite, testSimpleJiebaSearchModeTokenizer); |
| SUITE_ADD_TEST(suite, testSimpleJiebaAllModeTokenizer2); |
| SUITE_ADD_TEST(suite, testSimpleJiebaSearchModeTokenizer2); |
| |
| return suite; |
| } |
| // EOF |
