Skip to main content.
home | support | download

Back to List Archive

Re: words indexed, but not found on query

From: Bill Moseley <moseley(at)not-real.hank.org>
Date: Wed Apr 11 2001 - 22:34:34 GMT
Hi,

I don't have any answers for this exact problem, but if you upgrade to a
newer version you will probably get better support, not to mention phrase
searches, and faster indexing and searches.

http://sunsite.berkeley.edu:4444/swish-daily/ has the current development
code which seems to be stable at this point.

At 02:29 PM 04/11/01 -0700, Myke Komarnitsky wrote:
>I did search the archives, but couldn't find any pointers to a solution.
>
>I have indexed a web site using the file system method, and created an 
>index file.  If I test that index by doing:
>
>$ swish-e -D search.swish | grep "finger"
>finger: 17 4 1 1 23 8 1 1 65 8 9 1 86 4 9 1 448 3 9 1 449 4 9 1 450 4 9 1 
>451 4 9 1 452 4 9 1 453 4 9 1 454 3 9 1 455 4 9 1 456 4 9 1 457 4 9 1 458 4 
>9 1 459 4 9 1 460 4 9 1 461 3 9 1 462 5 9 1 463 4 9 1 464 4 9 1 465 4 9 1 
>466 2 9 1 467 2 9 1 468 4 9 1 469 3 9 1 470 3 9 1 471 3 9 1 472 4 9 1 473 3 
>9 1 474 3 9 1 475 4 9 1 476 4 9 1 477 4 9 1 478 2 9 1 479 2 9 1 480 4 9 1 
>481 4 9 1
>
>so obviously the word's in there.  but if I run a normal query, I get the 
>below:
>
>$ swish-e -f search.swish -w "finger"
># SWISH format 1.3
># Swish-e format 1.3
>#
># Name: Climbing Boulder
># Saved as: search.swish
># Counts: 152228800 words, 719 files
># Indexed on: 11/04/01 14:00:51 MDT
># Description: This is an index for climbingboulder.com
># Pointer: http://www.climbingboulder.com
># Maintained by: Myke Komarnitsky (myke@komar.org
># DocumentProperties: Enabled
># Stemming Applied: 152228848
># Search words: finger
>err: no results
>.
>
>what's especially odd is the line telling saying 150,000,000 words and 719 
>files. the file number is right, but there is DEFINITELY not anywhere close 
>to that many words.  any thoughts would be greatly appreciated.
>
>Thank you,
>Myke
>
>
>Michael Komarnitsky      Komar Consulting Group
>303.818.3718             http://consulting.komar.org
>http://climbingboulder.com
>-
>
>

Bill Moseley
mailto:moseley@hank.org
Received on Wed Apr 11 22:35:45 2001