How to Perform Searching Data by Using a Full-Text Search

After the full-text index has been created on a table, you can query the table by using the full-text predicates. The full-text predicates are used to specify how the search string should be searched in the table.
Predicates can be used to enhance the search so that the result have so focus on the search criteria. Sql programming have a list of predicates to be used.

The following predicates can be used while performing the full-text search:

FREETEXT: When the search criteria are given, FREETEXT searches for any variation of a word or a group of words given in the search column. FREETEXT is used for the prefix searches.
Considering the previous scenario of the bike racing competition, you can use the FREETEXT predicate to obtain the desired output, as shown in the following statement:

SELECT Description FROM Production.ProductDescription WHERE FREETEXT (Description, ‘race winners’)

CONTAINS: This predicate is used in queries when you want to search for a specific phrase or for the exact match. It also searches for the proximity of words within a text. For example, you can use the following statement to search for the words ‘Ride’ near the word ‘Bike’ in the ProductDescription table:

SELECT Description FROM Production.ProductDescription
WHERE CONTAINS (Desctiption, ‘ride NEAR bike’)

Populating the Full-Text Index in Search: SQL Server

After creating the full-text index, you need to populate it with the data in the columns enabled for full-text support. The SQL Server full-text search engine populates the full-text index through a process called population. Population involves filling the index with words and their location in the data page. When a full-text index is created, it is populated by default. In addition, the SQL Server automatically updates the full-text index as the data is modified in the associated tables.

However, the SQL Server does not keep a list of changes made to the indexed data when the CHANGE_TRACKING option is off. This option is specified while creating the full-text index by using the CREATE FULLTEXT INDEX statement.

If you do not want the full-text index to be populated when it is created using the CREATE FULLTEXT INDEX statement, then you must specify NO POPULATION along with the CHANGE TRACKING OFF option. To populate the index, you need to execute the ALTER FULLTEXT INDEX command along with the START FULL, INCREMENTAL, or UPDATE POPULATION clause.

For example, to create an empty full-text index on the ProductDescription table, you can execute the following statement:

CREATE FULLTEXT INDEX ON Production. ProductionDescription (Description)
KEY INDEX PK_ProductDescription_ProductDescriptionID
WITH CHANGE_TRACKING OFF, NO POPULATION

To populate the index you need to execute the following statement:
ALTER FULLTEXT INDEX ON Production. ProductDescription START FULL POPULATION

The preceding statement will populate the full-text index created on the ProductDesctiption table.

Similar to regular SQL indexes, full-text indexes can also be updated automatically as the data is modified in the associated tables. This repopulation can be time-consuming and adversely affect the usage of resources of the database server during periods of high database activity. Therefore, it is better to schedule repopulation of full-text indexes during periods of low database activity. You can specify the following types of full-text index population methods to repopulate the index:

Full Population

You can use this method when you need to populate the full-text catalog or the full-text index for the first time. After that, you can the maintain the indexes by using change tracking or incremental populations.
During a full population of a full-text catalo, index entries are built for all the rows in all the tables covered by the catalog. If a full population is requested for a table, index entries are built for all the rows in that table.

Change Tracking-Based Population

The SQL Server maintains a record of the rows that have been modified in a table set up for full-text indexing. These changes are propagated to the full-text index.

Incremental Timestamp-Based Population

The incremental population method updates the full-text index with the data that has been changed since the last time the index was refreshed. For an incremental population refresh to work, the indexed table must have a column of the timestamp data type. If a table does not have a column of the timestamp data type, then only a full population refresh can be done.

Configuring Full-Text Search in SQL Server

The full-text query feature in the SQL Server enables users to search for a wide range of text in the SQL tables. Consider an example. The sales management team of AdventureWorks, Inc. makes frequent searches on the ProductDescription table to develop marketing strategies. The search is based on the data stored in the Description column of the table.

A bike racing competition is scheduled to begin in Texas. The sales manager of AdventureWorks wants to see the details of all the bikes that are related to racing, so that a marketing strategy can be designed to increase the sale of these bikes. Specifically, he wants a list of all the bikes that have the keyword ‘race winner’ in the description.

As the data is large, the search query takes a long time to retrieve data from the table. In this scenario, you can apply a full-text index on the Description column of the ProductDescription table to improve the speed of searching.

To retrieve the required details by using full-text search, you need to configure full-text search on the database. For this, you need to perform the following tasks:

• Enable the full-text search in the database.
• Create a full-text catalog.
• Create a unique index.
• Create a full-text index.
• Populate the full-text index.

Enabling the Full-Text Search in the Database

Before using the full text search feature of the SQL Server, you need to enable the database using the following statement:
USE AdventureWorks
GO
Sp_fulltext_database enable
GO

Creating a Full-Text Catalog

A full-text catalog serves as a container to store full-text indexes. After enabling the full text search, you need to create a full-text catalog. A full-text catalog is a container that contains full-text indexes. A full-text catalog may have multiple full-text indexes. You can create a full-text catalog by using the following command:
CREATE FULLTEXT CATALOG cat1 AS DEFAULT

Creating a Unique Index

After creating the full-text catalog, you need to identify a unique index on the table. This unique index will be mapped to the values in the full-text index. You can use an existing unique index defined on the table, or create a new one. For example, you can create a unique index on the Production.ProductDescription table, as shown in the following statement:

CREATE UNIQUE INDEX Ix_Desc ON Production.ProductDescription
(ProductDescriptionID)

Creating a Full-Text Index

After you have created the full-text catalog and a unique index, you can create a full-text index on the table. A full-text index stores information about significant words and their location within a given column. You can use this information to compute full-text queries that search for rows with particular words or combinations of words. Full-text indexes can be crated on the base tables but not on the views or the system tables.

There are certain words that are used often and may hinder a query. These words are called noise words and are excluded from the search string. For example, if you search string is “Who is the governor of California”, a full-text search will not look for words, such as ‘is’ and ‘the’. Some noise words are a, an, the, and are.

Based on the preceding scenario, you can create a full-text index on the ProductionDescription table. This index is based on the lx_desc index created earlier on the Description column of the table.
Note: You can also create full-text index in the Object Explorer window by right-clicking the table, on which you need to create the full-text index, and selecting Full-Text index-Define Full-Text Index.