verity search results return application.cfm

[kejones]

We've got a couple of sites using verity collections to search cfml, html, and some mime documents. This is generally successful. However, we do see coldfusion script-only pages returning in the search results, including application.cfm, and other cfm pages that do not display content to the end-user. Since there is no HTML renderable content in these pages, why do they get included in the results?

How can I control this behavior? I've looked at the verity documentation, and am trying to use the mkvdk utility to delete application.cfm from the collection documents, but am getting an error. If I can get the syntax down correctly, is this the proper way of addressing the problem?

For instance, I am using the command string:
mkvdk -delete -collection C:\CFusionMX\verity\collections\intranet\file application.cfm
on my local development machine to try and test this, but get an error BadKeys: application.cfm

I don't understand why these types of files get included in the collection in the first place. Oh, we are running CFMX6.1.

Thanks for any help.
Kris

[universal_soldier]

<cfif (findnocase("htm", url))>output the result</cfif> works for us when we parse the results.

[kejones]

The issue is not how to remove all CFM files from the collection. We just want CFM files that do not output anything to the user to be excluded from the search index. Most of the files in this site are CFM files. Most importantly would be application.cfm, form action pages, and other included files.

Certainly we could parse the results for a filename in a predefined list, but we are trying to avoid defining this programmatically; thinking that it is odd that these files get included in the results at all. In the past we have segmented sites using .cfm and .cfml as a way of excluding specific files from the index (in the collection definition, we would not include the specific extension as required). But this site has been in production for some time, and is just now adding a search feature. We don't want to rename files, and have to recheck every single page and link in the entire site.

Does that explain what we're trying to do better?

Cheers,
Kris

[kejones]

Okay, so since no workable answers have come forth, we're adding a table of files to exclude from the results, and doing a query of queries against the results and the exclusion query, like this:

select *
from getresults
where right([key],charindex('\',reverse([key]),1)-1) not in (#quotedvaluelist(getexcluded.file_name)#)

However, we're getting an error that suggests that we cannot use the SQL right() function in a query of queries. Could this be true? I've tried a simple right([key],1)='m' and found that to also die with the same error. Any suggestions?

Cheers,
Kris

[daswerve]

This has always worked for me

[daswerve]

Sorry, forgot to add this (above the code in the last post):

[daswerve]

And add this (sorry, it's late):

[kejones]

Our list of files to not display from the search is too large to do a list exclusion (we've got folders of includes that were being returned in the results). Too bad that we have to add an IF statement inside the output loop, but that's the way the cookie crumbles I guess.

What we ended up doing is checking to see if the document had a title (which is returned in the verity resultset). Then we just need to make sure that any files we didn't want to provide in the returned results don't have a <title> attribute (or have an empty one).