Saturday, March 25, 2006

Numbers in SharePoint Search and noise filtering

Whenever you try to search in SharePoint and you are trying to search for a number combination - e.g. you have a custom metadata field contain productid in format P123-980 - you will notice that the search will not return the results you are expecting. The reason why this happens is that SharePoint uses a couple of files containing noise words which are omitted from the search - this will remove things such as "the", "a", ... - numbers are also added in there. What you need to do:
  • Remove the numbers from noiseenu.txt which you will find at [Sharepoint Portal Server Install directory]/Applications/Guid/Config

  • Restart the SharePoint search service

  • Reset the indexes and perform a full build

  • For more info check out How to customize SharePoint Portal Server 2003 by using IFilters, noise words, and thesaurus files and also take a look at how SharePoint Search works in multilinguage scenarios

    1 comment:

    decatec said...

    Thanks, was very useful in tuning MOSS Search