<![CDATA[PunBB Forums — Purpose of search tables]]> https://punbb.informer.com/forums/topic/3779/purpose-of-search-tables/ Sat, 04 Jun 2005 13:09:23 +0000 PunBB <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/44832/#p44832 just empty the tables then use maintenance in admin to regenerate them

]]> Sat, 04 Jun 2005 13:09:23 +0000 https://punbb.informer.com/forums/post/44832/#p44832 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/44831/#p44831 I'm currently transferring my forum from one database to another one. But the search_matches is quite big at the moment so I have some problems whith the transfer.

So is it 'ok' to put some values in the table but not all of them?

]]> Sat, 04 Jun 2005 12:54:17 +0000 https://punbb.informer.com/forums/post/44831/#p44831 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/14015/#p14015 Mediator wrote:

you could always cheat, and force the server to let you run it a long time....I use ini_set("max_execution_time","999999"); all the time.....then again most people don't have a 12 gig mysql of md5 hashes, and their matching words

Most hosts disable the ability to set the execution time during runtime. I think safe_mode even disables it.

]]> Wed, 19 May 2004 09:11:08 +0000 https://punbb.informer.com/forums/post/14015/#p14015 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/14002/#p14002 you could always cheat, and force the server to let you run it a long time....I use ini_set("max_execution_time","999999"); all the time.....then again most people don't have a 12 gig mysql of md5 hashes, and their matching words

]]> Wed, 19 May 2004 01:51:29 +0000 https://punbb.informer.com/forums/post/14002/#p14002 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13979/#p13979 But the query to generate that list can take a very long time to execute. I recently ran it on the forum database at sweclockers.com and I believe it took over a minute. By then, the script has terminated.

]]> Tue, 18 May 2004 15:05:37 +0000 https://punbb.informer.com/forums/post/13979/#p13979 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13973/#p13973 Yes, that's an alternative. Though making changes to the stopword list would be greatly facilitated by having available a most-commonly-used wordlist generated from the forum database. If the forum administrator does not have access to a tool like phpmyadmin and isn't comfortable writing his/her own query script, there's no way of generating that commonly-used wordlist.

]]> Tue, 18 May 2004 14:02:04 +0000 https://punbb.informer.com/forums/post/13973/#p13973 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13962/#p13962 An alternative would be to just inform the administrator of the existence of the stopwords list and suggest that he/she review it and make any changes they feel suitable for that perticular forum.

]]> Tue, 18 May 2004 12:20:38 +0000 https://punbb.informer.com/forums/post/13962/#p13962 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13955/#p13955 I agree. "Good" and "time" might be very significant words depending on the forum context -- for example, in a forum about racing.

Thinking about this, it seems that what is or isn't a stopword is quite dependent on the nature of the forum. What is a stopword for one forum may not be for another.

Which suggests the idea of a script to allow forum administrators to customize his/her stopwords; either a standalone script or something incorporated into the admin panel. I'd see it doing something like this:

1. On first run, briefly explain the concept of stopwords and why a forum administrator may want to customize the list;
2. On clicking a "Next" button, produce a list of common words (like the query above), except that the output would be loaded in a form, with each word a checkbox item;
3. The admin would checkmark some common words, and click a "Add to stopword list" button at page bottom.

]]> Tue, 18 May 2004 10:16:35 +0000 https://punbb.informer.com/forums/post/13955/#p13955 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13909/#p13909 sleddog: I'll revise the list for 1.2. Not sure all of those words are stopwords though. I'm primarily thinking of "good" and "time".

]]> Mon, 17 May 2004 14:40:28 +0000 https://punbb.informer.com/forums/post/13909/#p13909 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13899/#p13899 I few more candidate stopwords -- these appeared in my common words and are also included in the two lists above:

now well good time back still sure really

]]> Mon, 17 May 2004 09:42:20 +0000 https://punbb.informer.com/forums/post/13899/#p13899 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13896/#p13896 "fucking" is quite high on all forum word lists, it's remarkable how rude people on forums can get ... ah, fuck it! wink

]]> Mon, 17 May 2004 08:16:23 +0000 https://punbb.informer.com/forums/post/13896/#p13896 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13848/#p13848 Ah jesus......I am such a newb, never even noticed that function.....anyways here are the new results from the search

word  hits  
good 5370 
now 5056 
think 5026 
work 4956 
well 4781 
need 4693 
post 4568 
help 4179 
time 3868 
ill 3745 
people 3676 
thats 3613 
new 3275 
try 3157 
doesnt 2859 
still 2849 
right 2823 
really 2712 
com 2709 
back 2702 
because 2697 
say 2598 
program 2572 
first 2542 
someone 2535 
made 2534 
works 2435 
sure 2381 
add 2357 
anyone 2337 
ive 2281 
name 2229 
nice 2143 
again 2127 
same 2095 
shit 2089 
banned 2074 
look 2056 
start 2040 
give 2027 
put 2014 
tell 2005 
using 1972 
working 1912 
yeah 1900 
wont 1880 
fucking 1870 
stuff 1824 
maybe 1823 
read 1804 
trying 1777 
change 1761 
used 1736 
done 1731 
site 1695 
says 1694 
man 1690 
fuck 1682 
anything 1674 
fix 1671 
msn 1636 
wrong 1633 
edit 1630 
while 1604 
better 1599 
little 1526 
problem 1525 
sorry 1454 
down 1431 
long 1422 
mean 1407 
found 1404 
mine 1403 
check 1402 
plz 1397 
already 1368 
few 1359 
might 1358 
free 1354
]]> Sat, 15 May 2004 21:36:20 +0000 https://punbb.informer.com/forums/post/13848/#p13848 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13817/#p13817 Mediator: Great. I see a few more words that could be added to the list in 1.2.

Frank H: You're correct :)

]]> Sat, 15 May 2004 10:30:52 +0000 https://punbb.informer.com/forums/post/13817/#p13817 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13811/#p13811 wow, I see similarities ...
'think', 'thats' & 'ill' is very high for you too smile

hmm... no need to backup the search indexes ... as you can rebuild them in the forum admin panel, it takes time, but you have time for a cofeebreak (I did it myself after adding a few new stopwords). (I hope I'm correct, othervise Rickard probably see this wink)

]]> Sat, 15 May 2004 09:44:36 +0000 https://punbb.informer.com/forums/post/13811/#p13811 <![CDATA[Re: Purpose of search tables]]> https://punbb.informer.com/forums/post/13802/#p13802 
word  hits  
script 1442
good 1437
well 1318
post 1315
now 1231
think 1199
people 1091
need 1067
work 1039
ill 972
thats 939
scar 937
help 894
time 866
new 838
rs2 812
com 787
try 782
back 737
still 709
right 701
say 693
someone 654
because 653
first 653
really 652
shit 629
made 621
give 598
add 593
sure 588
doesnt 588
name 583
program 554
nice 551
start 549
banned 547
fucking 546
works 541
scripts 536
anyone 523
fuck 500
msn 496
stuff 491
read 491
yeah 485
ive 482
again 481
look 480
wont 477

Ok......well my forums has 3030 users and 76693 posts.....the results of the search where a bit scewed as I had to empty the search tables a few weeks ago as they where killing my backups.

Anyways......thats what I got

]]> Sat, 15 May 2004 05:03:40 +0000 https://punbb.informer.com/forums/post/13802/#p13802