使用全文搜索

SELECT SERVERPROPERTY(‘IsFullTextInstalled‘);

EXEC sys.sp_help_fulltext_system_components ‘filter‘;
或者
SELECT document_type, path
FROM sys.fulltext_document_types;

EXEC sys.sp_fulltext_service ‘load_os_resources‘, 1;

SELECT  lcid ,
        name
FROM    sys.fulltext_languages
ORDER BY name;

USE AdventureWorks2012 ;
GO
EXEC sp_configure ‘show advanced options‘, 1 ;
GO
RECONFIGURE
GO
EXEC sp_configure ‘default full-text language‘, 1043 ;
GO
RECONFIGURE
GO

SELECT stoplist_id, names
FROM sys.fulltext_stoplists;
SELECT stoplist_id, stopword, language
FROM sys.fulltext_stopwords;

SQL_Server_install_path\Microsoft SQL Server\MSSQL11.MSSQLSERVER\MSSQL\FTDATA\

EXEC sys.sp_fulltext_load_thesaurus_file 1033

--Table for documents
CREATE TABLE dbo.Documents
(
  id INT IDENTITY(1,1) NOT NULL,
  title NVARCHAR(100) NOT NULL,
  doctype NCHAR(4) NOT NULL,
  docexcerpt NVARCHAR(1000) NOT NULL,
  doccontent VARBINARY(MAX) NOT NULL,
  CONSTRAINT PK_Documents 
   PRIMARY KEY CLUSTERED(id)
);
GO

--Insert data
INSERT INTO dbo.Documents
(title, doctype, docexcerpt, doccontent)
SELECT N‘Columnstore Indices and Batch Processing‘, 
 N‘docx‘,
 N‘You should use a columnstore index on your fact tables,
   putting all columns of a fact table in a columnstore index. 
   In addition to fact tables, very large dimensions could benefit 
   from columnstore indices as well. 
   Do not use columnstore indices for small dimensions. ‘,
 bulkcolumn
FROM OPENROWSET(BULK ‘C:\TK461\ColumnstoreIndicesAndBatchProcessing.docx‘, 
                SINGLE_BLOB) AS doc;
INSERT INTO dbo.Documents
(title, doctype, docexcerpt, doccontent)
SELECT N‘Introduction to Data Mining‘, 
 N‘docx‘,
 N‘Using Data Mining is becoming more a necessity for every company 
   and not an advantage of some rare companies anymore. ‘,
 bulkcolumn
FROM OPENROWSET(BULK ‘C:\TK461\IntroductionToDataMining.docx‘, 
                SINGLE_BLOB) AS doc;
INSERT INTO dbo.Documents
(title, doctype, docexcerpt, doccontent)
SELECT N‘Why Is Bleeding Edge a Different Conference‘, 
 N‘docx‘,
 N‘During high level presentations attendees encounter many questions. 
   For the third year, we are continuing with the breakfast Q&A session. 
   It is very popular, and for two years now, 
   we could not accommodate enough time for all questions and discussions! ‘,
 bulkcolumn
FROM OPENROWSET(BULK ‘C:\TK461\WhyIsBleedingEdgeADifferentConference.docx‘, 
                SINGLE_BLOB) AS doc;
INSERT INTO dbo.Documents
(title, doctype, docexcerpt, doccontent)
SELECT N‘Additivity of Measures‘, 
 N‘docx‘,
 N‘Additivity of measures is not exactly a data warehouse design problem. 
   However, you have to realize which aggregate functions you will use 
   in reports for which measure, and which aggregate functions 
   you will use when aggregating over which dimension.‘,
 bulkcolumn
FROM OPENROWSET(BULK ‘C:\TK461\AdditivityOfMeasures.docx‘, 
                SINGLE_BLOB) AS doc;
GO

-- Search property list
CREATE SEARCH PROPERTY LIST WordSearchPropertyList;
GO
ALTER SEARCH PROPERTY LIST WordSearchPropertyList
 ADD ‘Authors‘ 
 WITH (PROPERTY_SET_GUID = ‘F29F85E0-4FF9-1068-AB91-08002B27B3D9‘, 
       PROPERTY_INT_ID = 4, 
       PROPERTY_DESCRIPTION = ‘System.Authors - authors of a given item.‘);
GO

-- Stopwords list
CREATE FULLTEXT STOPLIST SQLStopList;
GO
ALTER FULLTEXT STOPLIST SQLStopList
 ADD ‘SQL‘ LANGUAGE ‘English‘;
GO

--  Check the Stopwords list
SELECT w.stoplist_id,
 l.name,
 w.stopword,
 w.language
FROM sys.fulltext_stopwords AS w
 INNER JOIN sys.fulltext_stoplists AS l
  ON w.stoplist_id = l.stoplist_id;
GO

--第一个参数是用来分析的字符串，第二个是语言ID，第三个是stoplist ID （用前面的语句查询），第四个是否解析重音
SELECT *
FROM sys.dm_fts_parser
(N‘"Additivity of measures is not exactly a data warehouse design problem.
However, you have to realize which aggregate functions you will use
in reports for which measure, and which aggregate functions
you will use when aggregating over which dimension."‘, 1033, 5, 0);
SELECT *
FROM sys.dm_fts_parser
(‘FORMSOF(INFLECTIONAL,‘+ ‘function‘ + ‘)‘, 1033, 5, 0);

SELECT *
FROM sys.fulltext_semantic_language_statistics_database;

CREATE DATABASE semanticsdb ON
(FILENAME = ‘C:\Program Files\Microsoft Semantic Language Database\semanticsdb.mdf‘),
(FILENAME = ‘C:\Program Files\Microsoft Semantic Language Database\semanticsdb_log.ldf‘)
FOR ATTACH;

EXEC sp_fulltext_semantic_register_language_statistics_db
 @dbname = N‘semanticsdb‘;
GO

CREATE FULLTEXT CATALOG DocumentsFtCatalog;

-- Full-text index
CREATE FULLTEXT INDEX ON dbo.Documents
( 
  docexcerpt Language 1033, 
  doccontent TYPE COLUMN doctype
  Language 1033
  STATISTICAL_SEMANTICS
)
KEY INDEX PK_Documents
ON DocumentsFtCatalog
WITH STOPLIST = SQLStopList, 
     SEARCH PROPERTY LIST = WordSearchPropertyList, 
     CHANGE_TRACKING AUTO;
GO

SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘data‘);

--  Logical operators - OR
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘data OR index‘);

-- Logical operators - AND NOT
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘data AND NOT mining‘);

-- Logical operators - parentheses
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘data OR (fact AND warehouse)‘);

-- Phrase
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘"data warehouse"‘);

--  Prefix
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘"add*"‘);

--  Simple proximity
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘NEAR(problem, data)‘);

-- Proximity with max distance
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘NEAR((problem, data),5)‘);
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘NEAR((problem, data),1)‘);

--  Proximity with max distance and order
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘NEAR((problem, data),5, TRUE)‘);

-- Inflectional forms
-- The next query does not return any rows
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘presentation‘);
-- The next query returns a row
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘FORMSOF(INFLECTIONAL, presentation)‘);
GO

<XML ID="Microsoft Search Thesaurus">
    <thesaurus xmlns="x-schema:tsSchema.xml">
    <diacritics_sensitive>0</diacritics_sensitive>
        <expansion>
            <sub>Internet Explorer</sub>
            <sub>IE</sub>
            <sub>IE5</sub>
        </expansion>
        <replacement>
            <pat>NT5</pat>
            <pat>W2K</pat>
            <sub>Windows 2000</sub>
        </replacement>
        <expansion>
            <sub>run</sub>
            <sub>jog</sub>
        </expansion>
        <expansion>
            <sub>need</sub>
            <sub>necessity</sub>
        </expansion>
    </thesaurus>
</XML>

EXEC sys.sp_fulltext_load_thesaurus_file 1033;

--  Synonyms
-- The next query does not return any rows
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘need‘);
-- The next query returns a row
SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(docexcerpt, N‘FORMSOF(THESAURUS, need)‘);

SELECT id, title, docexcerpt
FROM dbo.Documents
WHERE CONTAINS(PROPERTY(doccontent,‘Authors‘), ‘Dejan‘);

SELECT id, title, doctype, docexcerpt
FROM dbo.Documents
WHERE FREETEXT(docexcerpt, N‘data presentation need‘);

--   Rank with CONTAINSTABLE
SELECT D.id, D.title, CT.[RANK], D.docexcerpt
FROM CONTAINSTABLE(dbo.Documents, docexcerpt, 
      N‘data OR level‘) AS CT
 INNER JOIN dbo.Documents AS D
  ON CT.[KEY] = D.id
ORDER BY CT.[RANK] DESC;

--  Rank with FREETEXTTABLE
SELECT D.id, D.title, FT.[RANK], D.docexcerpt
FROM FREETEXTTABLE (dbo.Documents, docexcerpt, 
      N‘data level‘) AS FT
 INNER JOIN dbo.Documents AS D
  ON FT.[KEY] = D.id
ORDER BY FT.[RANK] DESC;

--  Weighted terms
SELECT D.id, D.title, CT.[RANK], D.docexcerpt
FROM CONTAINSTABLE
      (dbo.Documents, docexcerpt, 
       N‘ISABOUT(data weight(0.8), level weight(0.2))‘) AS CT
 INNER JOIN dbo.Documents AS D
  ON CT.[KEY] = D.id
ORDER BY CT.[RANK] DESC;

--  Proximity term
SELECT D.id, D.title, CT.[RANK]
FROM CONTAINSTABLE (dbo.Documents, doccontent, 
      N‘NEAR((data, row), 30)‘) AS CT
 INNER JOIN dbo.Documents AS D
  ON CT.[KEY] = D.id
ORDER BY CT.[RANK] DESC;

-- 1. Top 20 semantic key phrases
SELECT TOP (20)
 D.id, D.title, SKT.keyphrase, SKT.score
FROM SEMANTICKEYPHRASETABLE
      (dbo.Documents, doccontent) AS SKT
 INNER JOIN dbo.Documents AS D
  ON SKT.document_key = D.id
ORDER BY SKT.score DESC;

-- 2. Documents that are similar to document 1
SELECT SST.matched_document_key, 
 D.title, SST.score
FROM SEMANTICSIMILARITYTABLE
     (dbo.Documents, doccontent, 1) AS SST
 INNER JOIN dbo.Documents AS D
  ON SST.matched_document_key = D.id
ORDER BY SST.score DESC;

-- 3. Semantic search key phrases that are common to two documents
SELECT SSDT.keyphrase, SSDT.score
FROM SEMANTICSIMILARITYDETAILSTABLE
      (dbo.Documents, doccontent, 1,
       doccontent, 4) AS SSDT
ORDER BY SSDT.score DESC;
GO

-- 4. Clean up
DROP TABLE dbo.Documents;
DROP FULLTEXT CATALOG DocumentsFtCatalog;
DROP SEARCH PROPERTY LIST WordSearchPropertyList;
DROP FULLTEXT STOPLIST SQLStopList;
GO