Db2 Text Search SCORE function

The score of a document is dynamic and calculated independently for each query.

Updates to a document as well as adding or removing documents from a text index can cause a change of the score of a document for a query term.

Assume there is a set of documents discussing transportation and pollution. If you want to locate documents containing references to both terms, but only if the term pollution scores higher than the term transportation, you can use the following command:

	SELECT document_id 
	FROM document_library
	WHERE SCORE(document_content, 'pollution') > 
	SCORE(document_content, 'transportation')
	and CONTAINS(document_content, 'transportation pollution') = 1

To enhance performance, you can format your query to use the boost (^) modifier so that the search function is run only once, as follows:

	SELECT document_id 
	FROM document_library
	WHERE SCORE(document_content, 'pollution^10 transportation') > 0 
	ORDER BY SCORE(document_content, 'pollution^10 transportation') DESC

The first query does not return any results if pollution scores low. The second query gives higher importance to pollution but still returns documents if pollution scores low in all documents.