Thursday, November 29, 2012

Hit Highlighting and Summarization with SQL Server Full Text Indexing/Searching? Here you go...

Bob Beauchemin's Blog - Hit Highlighting/Summarization product for SQL Server Full-Text Search

"I've often asked about add-ins for SQL Server Full-Text Search feature. And the most requested feature for FTS has long been some way to accommodate hit highlighting. Recently, I've come across a beta for a product that does just this, along with document summarization formats for display, ThinkHighlight by Interactive Thoughts.

The add-in is implemented as a single SQLCLR assembly that exposes a pair of UDFs, HitHighlightContext and HitHighlight. The first UDF prepares a context. You only have to do this once per query, so this appears to be a nice optimization. The HitHighlight one is used to produce a column of output with summarized text and highlighted hits. You have a choice of three different output summarization strategies: baseline, complete, and top-fragment. Baseline and top-fragment are meant to be used in a manner similar to one that you'd see used in search websites. There's an interactive demo that shows some examples.

The product comes with a ..."

Interactive Thoughts - ThinkHighlight for SQL Server


Indexing and searching documents with Microsoft's SQL Server Full-Text Search is quite easy and effective, yet visualizing the results has always needed a UI boost.

Introducing ThinkHighlight for SQL Server, a robust plugin component that allows users a much better understanding of the results in relation to their query.

Key Features

  • Hit-highlighting of query terms in the results
  • Snippet generation with several summarization strategies
  • Works directly on both text and binary columns (most IFilters are supported)
  • Multilingual support in the document/index/query level
  • Linked highlighting, allows styling per query term/group

More Information


That is something that's always concerned me about SQL Server's Full Text Indexing (which has gotten really awesome in SQL Server 2012), the apparent lack of third party extensions, ISV's pushing it, etc.

This looks very interesting and when doing searching, kind of a must have...

No comments: