Delete URLs manually when all fails

edev · Post by **edev** » Mon Jan 15, 2007 11:20 am

Hi,

I have a profile as a base, and I combined another database with this profile using a script like this:
<DB=$sourceDB>
<SQL ROW "select * from html">
<DB=$targetDB>
<SQL NOVARS "insert into html values ($id, $Hash, $Size, $Visited, $Dlsecs, $Depth, $Url, $Title, $Body, $Keywords, $Description,
$Meta, $Catno, $Modified, $NextCheck, $Views, $Clicks, $CTR, $Pop, $MimeType, $Charset)">
</SQL>
</SQL>
</A>

Then I did a reindex so I would get a combination of two databases. The problem is I cannot delete URLs from the Edit/Delete URL option on the admin page. I want to delete all URLs matching a pattern like http://www.somesite.com/PrinterFriendly*, the goal is to get rid of any URL matching *PrinterFriendly*.
can I do this using a delete statement like
"texis -d E:\profile/db2 - s delete from html where URL like '*PrinterFriendly*'"?
Can this be done in a command window, or do I have a write a script to do this? Thanks!

Post by **mark** » Mon Jan 15, 2007 11:52 am

Why can't you use list/edit urls to delete?

For simple sql statements you can execute them from the command line or in a script. You could use one of these
... where Url like 'printerfriendly'
... where Url matches '%PrinterFriendly%'
... where Url matches 'http://www.somesite.com/PrinterFriendly%'

edev · Post by **edev** » Mon Jan 15, 2007 12:59 pm

Thank you John. When I try to delete the URLs from the admin screen I get this:

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\error.tbl: No file write permission in the function kdbf_free

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr: No file write permission in the function kdbf_put

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr: No file write permission in the function kdbf_put

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr: No file write permission in the function kdbf_put

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorid.btr: No file write permission in the function kdbf_put

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorid.btr: No file write permission in the function kdbf_put

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorid.btr: No file write permission in the function kdbf_put

006 /webinator/dowalk(logerror) 2428: Cannot delete value (http://www.thecanadianencyclopedia.com/ ... 1ARTFET_E8) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr

006 /webinator/dowalk(logerror) 2428: Cannot delete value (45802f29630) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorid.btr

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\error.tbl: No file write permission in the function kdbf_free

006 /webinator/dowalk(logerror) 2428: Cannot delete value (http://www.thecanadianencyclopedia.com/ ... 1ARTFET_E8) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr

006 /webinator/dowalk(logerror) 2428: Cannot delete value (45802f29630) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorid.btr

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\error.tbl: No file write permission in the function kdbf_free

006 /webinator/dowalk(logerror) 2428: Cannot delete value (http://www.thecanadianencyclopedia.com/ ... 1ARTFET_E8) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr

006 /webinator/dowalk(logerror) 2428: Cannot delete value (45802f29630) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorid.btr

006 /webinator/dowalk(logerror) 2428: Can't write to KDBF file E:\CultureCaWalks\CultureCa_Site1b\db1\error.tbl: No file write permission in the function kdbf_free

006 /webinator/dowalk(logerror) 2428: Cannot delete value (http://www.thecanadianencyclopedia.com/ ... 1ARTFET_E8) from index E:\CultureCaWalks\CultureCa_Site1b\db1\xerrorurl.btr

edev · Post by **edev** » Mon Jan 15, 2007 1:34 pm

I tried to delete from a command line:
C:\Program Files\Thunderstone Software\Webinator>texis -d E:\CultureCaWalks\CultureCa_Site1b\db1 -s "delete from html where URL matches 'http://www.thecanadianencyclopedia.com/PrinterFriendly%'"
Texis Web Script (Vortex) Copyright (c) 1996-2004 Thunderstone - EPI, Inc.
Commercial Webinator Version 5.00.1085675754 20040527 (i686-intel-winnt-32-32)

115 Field URL non-existent
000 SQLPrepare() failed with -1 in the function prepntexis

...and got the above error. Help?

Post by **mark** » Mon Jan 15, 2007 2:18 pm

Your Webinator installation is seriously hosed and you're lucky if anything works. If you don't have perms to delete you certainly don't have them for crawling either.

John followed your example and called the field URL. It's called Url. Case matters.

edev · Post by **edev** » Mon Jan 15, 2007 2:28 pm

Hi Mark,

I know the installation was somewhat non-standard...I didn't install it and we have already started a new server in hope it would solve the problem. But in the mean time I'll have to work with this server...from what you looked at could you recommend anything in the installation? thanks...

Post by **mark** » Mon Jan 15, 2007 2:58 pm

Try making the database directory and all files in it full control for everyone. Make sure it's always accessed as the same user, keeping in mind that the webserver's CGI user is always accessing it for searches and web admin. Most likely anything you ran from the command line would have been as a different user and would either not work or would cause problems the web interface down the line.

edev · Post by **edev** » Mon Jan 15, 2007 3:03 pm

I have an internet user set up for web interface of Webinator, and I am the local admin of the server (Windows 2003 server) so anything I run on command will be under the local admin rights.

I am now running a query on command

texis -d E:\CultureCaWalks\CultureCa_Site1b\db1 -s "delete from html where Url matches 'http://www.thecanadianencyclopedia.com/PrinterFriendly%'"

and now the command seems to freeze after this:
Texis Web Script (Vortex) Copyright (c) 1996-2004 Thunderstone - EPI, Inc.
Commercial Webinator Version 5.00.1085675754 20040527 (i686-intel-winnt-32-32)

id Hash Size Visited Dlsecs Depth Url Title Body Keywords Description Meta Catno Modified NextCheck Views Clicks CTR Pop MimeType Charset
------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+------------+

It's been 30 minutes I'm still waiting for the results to come up. There are over 33,000 URLs that match this pattern and the database is about 420,000 Urls, is this command just taking too long?

Post by **mark** » Mon Jan 15, 2007 3:52 pm

A delete won't print anything but the headers. It could take some time to delete all of those, especially if the machine/database is in use by anything else, like search.

edev · Post by **edev** » Thu Jan 18, 2007 3:31 pm

Thanks Mark, the window froze for a long time, then I closed it, and did a select statement

texis -d E:\CultureCaWalks\CultureCa_Site1b\db1 -s "select Url from html where Url matches 'http://www.thecanadianencyclopedia.com%'"

and it seems that all Urls were deleted!