Page 1 of 1
Category walk problem
Posted: Fri Aug 29, 2003 2:07 pm
by mmcfadden
A search in everything will works great. I get the right results without fail so far. However I will then go to the advanced page and choose a dropdown category and I won't get all the results that I should. When I have my results from everything I will pick out a result that has the URL pattern that is in a category and I won't find that result in the category search results. I then take that URL pattern and enter it into the List/Edit URL utility and I get that document back correctly. So I can verify that my category patterns are correct and that the index does indeed include the document I would expect. It seems to be directly related to the category. This problem seems to occur when I run a new walk by hitting the go button. After the walk is done the Categories are hit and miss. Then after the walk is done and the Categories won't work I will just hit the update button which does a recategorization and everything seems to work at this point. I would really like to automate the walk on a weekly basis but if the categorization fails every time I will have to run a walk and then use a manual process of running the walk and then hitting the update button. Any ideas what is wrong here?
Category walk problem
Posted: Fri Aug 29, 2003 2:35 pm
by mark
What's an example category pattern and url that's missed during walk but not on recat?
If you look at the info about the url in question under list/edit urls does the category show there?
"Update" shouldn't do anything unless you've changed something. It shouldn't recat unless you change the categories.
What version of Webinator? Try downloading the latest dowalk and webinatoradmin if you haven't already.
Not that I'm proposing it as a solution, but you can invoke recategorization by hand. See "using dowalk" in the manual and look for recat.
Category walk problem
Posted: Fri Aug 29, 2003 3:08 pm
by mmcfadden
I am using the latest version 4.3.7
We have a pattern for a category that is */ADAAG/* so I just ran a search on toilet in everything and got back a document with this as part of its url /ADAAG/figures/figa7.html. I had to run a new walk to break the site again. After this I was only able to get this document in a search of everything and I didn't get any results in the category that has the */ADAAG/* pattern in it. I went into List/Edit URL and put in */ADAAG/figures/figa7.html which returned the document and looked in the category field and it said none. I then went out to All Walk Settings and clicked update (by the way I didn't make any changes to categories or settings I only hit update). I then went back in to List/Edit URL and entered the same pattern and found the the document /ADAAG/figures/figa7.html is now in the proper category. Finally I went to the advanced search page entered toilet as a search term and chose the Category with the */ADAAG/* pattern and now I get 33 results. I am not sure why this is going on any other ideas?
Category walk problem
Posted: Fri Aug 29, 2003 3:58 pm
by mark
Not really. Will have to experiment and try to replicate...
Category walk problem
Posted: Fri Aug 29, 2003 5:19 pm
by mark
It appears to be related to having ignore case on. It will require further investigation to fix... Watch this space...
Category walk problem
Posted: Fri Aug 29, 2003 6:08 pm
by mark
Workaround is to enter your categories in all lowercase when using ignore case.
Or download and install the 4.3.8 scripts from the webinator examples page.
Category walk problem
Posted: Fri Aug 29, 2003 6:27 pm
by mmcfadden
Was anything other than this problem updated in this release? The reason I am asking is that we have made some modifications to the search script and it would be easy to replace the dowalk and admin script but the search script will require some work. Thank you for your quick response. It is amazing how fast this was fixed.
Category walk problem
Posted: Fri Aug 29, 2003 10:55 pm
by mark
Search has not changed.
You can always see what has changed by looking at the change log at the bottom of the script itself.