Verity Knowledge - How do different search types work in Knowledge Management and Background Searching

Version 3

    In order to understand more about how verity searches work look at the Verity Query Language.pdf which is installed by default on your Verity server.

    In ServiceDesk Pre version 7.2.5 – It uses the ACCRUE search function both within Knowledge Management Search and Background Searching (only natural language search was available).

    After version 7.2.5 pre 7.3.2 – It uses the FREE TEXT function for natural language search and also for background searches. As the nature of background search is meant to be that you enter text as you would normally, it uses the FREE TEXT function (also known as QBE). This means that it automatically excludes words such as "the", "has", etc.. The whole idea of the background search is that you type in a text field as normal and then it performs a search on this.


    From 7.2.5. you also now have the option for selecting the STANDARD for the search option on the Knowledge Management component. The Standard option queries the verity collections in the same way as the verity dashboard test search does (although Verity searches will not include things like knowledge domains, classifications, and effectiveness scores).  Standard search allows you to be more specific about what you want to search on.


    So using standard search, here are some example syntax strings that you could search using:


    apple pie

    This would look for records in the verity collections which contain either the word apple or pie.  Then it would score the records based on how many times the word apple or the word pie appears in the record and put the top scoring record to the top of the search results.  So for example, a record with apple apple apple apple in the text would appear higher than a record with just apple pie in the text.  It is just looking for appearances of either word.


    apple AND pie

    This would just bring back your records which contains both of these words within the text.  Again, the records with the words apple and pie in them the more times would appear higher in the search results.


    "apple pie"

    This would just bring back your records which contains this exact phrase within the text in this specific order .  Again, the records with the exact phrase apple pie in them in this order the most times would appear higher in the search results.


    apple, [100] pie

    This gives an additional weighting to the word pie of 100.  This means that records that contain the word apple will still be retrieved by the search results but those with the word pie in will go straight to the top of the search results list.