My links sql database holds html in a field. I only want to index for the search results, and not the html markup. I could get around this by creating a new field, and insert into that a text only version of the field. However, this will double the size of the database - a very large size increase.

Is there a way to block certain words or block <tags> from being indexed in linksql?

If you're willing to a little bit of hacking, you can take a look at the admin/GT/SQL/Search/... directories for the modules that will handle the indexing. If you are using the internal indexing method, in the file admin/GT/SQL/Search/Base/Common.pm, the function tokenize will be of interest. (throw in some code that will strip out the html before tokenizing in there)

Otherwise, a quick but not perfect fix might be to put all the tag words into the stop word list. At the top of the admin/GT/SQL/Search/Base/Common.pm file, you'll find the $STOPWORDS array ref and put stuff like href, style, etc into it. You don't have to worry about words smaller than 3 as they are automatically ignored by the indexer.

Thanks for the helpful and quick response Aki. Very useful

Wow.... I feel so privileged, a "thank you" from the secretary of defense.

Good luck with either option
Aki

lol.

BTW I've probably just run the biggest test of the indexing system it's likely to get for a while (lots of text, lots of records) and it still performs really well. I wasn't expecting that, so congrats.

I was expecting to have to go and buy something else, but it looks as though links does the job extremely well.