All,



I've recently been working on the full text searching capabilities of IBO.
With the next release IBO will look for the existence of a certain stored
procedure that will allow for more accurate processing of phrase based
searching.



Previously I just used CONTAINING but it would require that the phrases
match exactly. If there was a comma or a hyphen or an apostrophe that wasn't
identical then the search would fail. However, I want the searching to be
tolerant of those types of minor differences. This new stored procedure will
actually break everything down and parse through things and perform a phrase
based comparison that is tolerant of punctuation differences.



The way IBO handles phrase based searching is it uses the full text indexes
to find the hits of all items containing the words, but they could be in any
order so that is only one step toward the finished result set. I used this
stored procedure to finalize the result set. It omits those instances where
even though all the words are there they are not in an actual phrase.



I also included in this stored procedure the ability to make use of
wildcards inside of the phrase. I'll give an example.



In your text you have the phrases "In the beginning" and "in the
beginnings". Let's say you want your search to pick up both instances. This
is now possible by entering search criteria like this "In the begin%".



I also made it so that you can use negative phrases. If you put a minus sign
right before it then it will push out all matches that contain that phrase.
This can help give you more precision to push out the noise to help you dig
for what you are actually after.



Jason



[Non-text portions of this message have been removed]