Re: Different wildcard/hyphen behaviour in Windows 2003 Server?

Tech Tip: Click here to run a free scan for Windows Errors and optimize PC performance

Yes, the English wordbreaker in Windows 2003 Server is different from the
one in Windows 2000. In most cases, query results are better now, but this
particular case looks somewhat broken. Win2k actually thinks 1-2-3 is a
date, but it works to its advantage!

Here is the output from the two wordbreakers. (the clsid is from
hklm\system\currentcontrolset\control\contentindex\language\english_us\wbreakerclass)

w2k3:

R:\>lrtest /b /c:{80A3E9B0-A246-11D3-BB8C-0090272FA362} "1-2-3"
dll loaded: U:\WINDOWS\system32\LangWrbk.dll
dll version 5.2.3790.0
file create time: 11-03-2003 11:35a
file last write time: 3-25-2003 5:00a
FileVersion: '5.2.3790.0 (srv03_rtm.030324-2048)'
FileDescription: 'English wordbreaker'
CompanyName: 'Microsoft Corporation'
ProductName: 'Microsoft« Windows« Operating System'
Wordbreaker requires license: No
Wordbreaker license: 'Copyright Microsoft Inc.'
Original text: '1-2-3'
IWordSink::PutAltWord: cwcSrcLen 1, cwcSrcPos 0, cwc 1, '1'
IWordSink::PutWord: cwcSrcLen 1, cwcSrcPos 0, cwc 3, 'NN1'
IWordSink::PutAltWord: cwcSrcLen 1, cwcSrcPos 2, cwc 1, '2'
IWordSink::PutWord: cwcSrcLen 1, cwcSrcPos 2, cwc 3, 'NN2'
IWordSink::PutAltWord: cwcSrcLen 1, cwcSrcPos 4, cwc 1, '3'
IWordSink::PutWord: cwcSrcLen 1, cwcSrcPos 4, cwc 3, 'NN3'

w2k:
C:\>lrtest /b /c:{59e09780-8099-101b-8df3-00000b65c3b5} "1-2-3"
dll loaded: C:\WINNT\system32\infosoft.dll
dll version 5.0.2195.6601
file create time: 7-09-2003 1:40p
file last write time: 6-19-2003 12:05p
FileVersion: '5.00.2195.6601'
FileDescription: 'Wordbreaker and stemmer dll'
CompanyName: 'Microsoft Corporation'
ProductName: 'Microsoft(R) Windows (R) 2000 Operating System'
Wordbreaker requires license: Yes
Wordbreaker license: 'Copyright Infosoft International Inc.'
Original text: '1-2-3'
IWordSink::PutAltWord: cwcSrcLen 5, cwcSrcPos 0, cwc 5, '1-2-3'
IWordSink::PutWord: cwcSrcLen 5, cwcSrcPos 0, cwc 10, 'DD19030102'





"Jeremy Scrivener" <jscriv@xxxxxxxxxxxx> wrote in message
news:1115328842.075310.281390@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> Hi All,
>
> The company I work for has a product based on Indexing Service.
>
> Since several of our clients have migrated to Windows 2003 Server, they
> are experiencing different results with queries based on wildcards and
> hyphens to that experienced on Windows 2000. The query itself resembles
>
> "@propertyname 1-2-3*".
>
> [NB - This is not a dialect issue as far as I can tell as the query is
> an "@propertyname" query]
>
> I have tried removing all the numbers from the noise.enu file,
> rescanned, but this did not help - in fact, this stopped results being
> returned on an exact match query! (& doubled the size of the catalog
> ;-)
>
> Can anyone confirm that behaviour has changed with 2003 Server, and if
> so provide any details? We're considering opening a PSS with MS on
> this, but you would hope that such details would be documented
> somewhere!
>
> Thanks in advance,
>
> Jeremy.
>


.

Relevant Pages

  • Re: Different wildcard/hyphen behaviour in Windows 2003 Server?
    ... looks like it isn't possible to load the Windows 2000 wordbreaker ... - in order to replace this dll with the one from Windows 2000, ...
    (microsoft.public.inetserver.indexserver)
  • Wordbreaker problems
    ... I'm trying to get proper wordbreaker compound noun decomposition behaviour ... query side and have a few question. ... If the wordbreaker does not put base components into the index, ...
    (microsoft.public.inetserver.indexserver)
  • Re: Fulltext not returning all rows
    ... I have just one word in the notes column. ... contains query. ... >Windows 2000 Server, Full-text Search is using the OS-supplied ... This wordbreaker has known issues with any ...
    (microsoft.public.sqlserver.fulltext)