SQL Server Full Text Search Noise Words and Thesaurus Configurations

By:   |   Comments   |   Related: > Full Text Search


I have heard that Full Text Search uses a noise words to eliminate meaningless words in searches.  I have also heard that a thesaurus is used, but I am not exactly sure how.  Can you provide some details related to how both of these technologies are used in Full Text Search?  I am interested in more of a background on these specific Full Text Search technologies as well as where the files are located and how I can update them.


Although the noise words and thesaurus have many similarities in terms of the implementation with Full Text Search, let's break apart each technology in order to explain them with more clarity.  Keep in mind that the information in this tip relates to SQL Server 2005, although SQL Server 2000 and 7.0 have some of the same properties.  Stay tuned for information on Full Text Search in SQL Server 2008 because some of the rules have changed.

Noise Words

  • Purpose: The noise words file has been established by Microsoft on a per language basis to determine the words which do not add value to the search.  For example, in the English language (United  States) noise words would include: because, been, before, being, between, both, but, etc.  In the Noiseenu.txt file (English US) over 100 noise words are setup by default and this file is used by default when querying with full text search.
  • Language Support: 20 different languages are supported from Chinese (Noisechs.txt) to Turkish (Noisetrk.txt).
  • Windows Directory: $SQL_Server_Install_Path\Microsoft SQL Server\MSSQL.1\MSSQL\FTDATA
  • Modifications: New noise words can be added to the file.  As we have had to make changes, we add them after the last entry which is 'your' by default.
  • Additional Information
    • If you make a change to a noise word file, it is necessary to repopulate all of the full text catalogs before any of the new noise words will be used.


  • Purpose: Ability to define synonyms and use the synonyms in full text searches.  Specifically with the FORMSOF THESAURUS parameter with the CONTAINS or CONTAINSTABLE commands the thesaurus is used to determine synonyms of the search terms.  The same is true with the FREETEXT and FREETEXTTABLE commands.  With these commands the thesaurus is used to identify expressions or replacements for the searched terms.
  • Language Support: The thesaurus files are in XML format with a global file (tsGLOBAL.xml) and then 18 language specific files.  By default all of the files have sample XML that is commented out.  So by default no synonyms are setup when a full text search is issued.
  • Windows Directory: $SQL_Server_Install_Path\Microsoft SQL Server\MSSQL.1\MSSQL\FTDATA
  • Modifications: After reviewing the XML format, two types of thesaurus entries exist..  First are the expansion set and second are the replacement set.  The expansion set which can be considered synonyms or substitute terms.  The replacement set indicates that one string is replaced with another.
  • Additional Information:
    • All thesaurus files should be saved as Unicode.
    • The Full Text Search Engine Windows Service must be restarted to begin to use the new thesaurus files.
Next Steps

sql server categories

sql server webinars

subscribe to mssqltips

sql server tutorials

sql server white papers

next tip

About the author
MSSQLTips author Jeremy Kadlec Jeremy Kadlec is a Co-Founder, Editor and Author at MSSQLTips.com with more than 300 contributions. He is also the CTO @ Edgewood Solutions and a six-time SQL Server MVP. Jeremy brings 20+ years of SQL Server DBA and Developer experience to the community after earning a bachelor's degree from SSU and master's from UMBC.

This author pledges the content of this article is based on professional experience and not AI generated.

View all my tips

Comments For This Article

get free sql tips
agree to terms