Hi,
Let's assume you're using Solr version 3.1.0 and an unmodified FieldType
"text_rev". It looks like this:
<fieldType name="text_rev" class="solr.TextField"
positionIncrementGap="100">
<analyzer type="index">
<tokenizer class="solr.WhitespaceTokenizerFactory"/>
<filter class="solr.StopFilterFactory" ignoreCase="true"
words="stopwords.txt" enablePositionIncrements="true" />
<filter class="solr.WordDelimiterFilterFactory" generateWordParts="1"
generateNumberParts="1" catenateWords="1" catenateNumbers="1" catenateAll="0"
splitOnCaseChange="0"/>
<filter class="solr.LowerCaseFilterFactory"/>
<filter class="solr.ReversedWildcardFilterFactory" withOriginal="true"
maxPosAsterisk="3" maxPosQuestion="2" maxFractionAsterisk="0.33"/>
</analyzer>
...
Also let's assume that what you have two docs in your index with these URLs:
A:"http://my.host/SPC265_SharePoint_2010.pptx";
B:"http://my.host/OpenTRs2010.xlsx";
Now you want to match only A and not B, and you attempt that using q=url:_2010
What happens here can easily be simulated by
http://localhost:8983/solr/admin/analysis.jsp:
Your Tokenizer keeps the whole URL as a token.
The WordDelimiterFilter splits on all kinds of things, also removing the "_".
Thus you get a match on 2010
What you need to do is design a new FieldType in your schema specifically for
your need.
Choose a Tokenizer based on what you want to be your tokens.
My suggestion is like this:
<fieldType name="urltype" class="solr.TextField" positionIncrementGap="100">
<analyzer>
<charFilter class="solr.PatternReplaceCharFilterFactory"
pattern="_" replacement=" UNDERSCORE " />
<tokenizer class="solr.PatternTokenizerFactory" pattern="[:/\\.\\? ]" />
</analyzer>
</fieldType>
Now your tokens will be "http my host SPC265 UNDERSCORE SharePoint UNDERSCORE
2010 pptx"
A search for url:_2010 would match because the _ is replaced with a special
token which can then be matched. Proof:
You could do similar thins for other special cases you wish to match. I assume
that the normal case is that you want to match whole words like sharepoint or
pptx, and that the _ matching is a special case.
--
Jan Høydahl, search solution architect
Cominvent AS - www.cominvent.com
Solr Training - www.solrtraining.com
On 14. juni 2011, at 11.42, Tirthankar Chatterjee wrote:
> Eric
> Thx for the reply. But what can I do to avoid getting 2010. I wanted a phrase
> query with underscore, so it would return results with underscore2010 only.
>
> Sent from iPod
>
> On Jun 13, 2011, at 3:47 PM, "Erick Erickson" <erickerick...@gmail.com> wrote:
>
>> You haven't supplied the information that's really
>> needed to help here, please review:
>>
>> http://wiki.apache.org/solr/UsingMailingLists
>>
>> But at a guess your analysis chain contains
>> WordDelimiterFilterFactory, which is splitting
>> the input stream into tokens on letter/number
>> changes, and capitalization changes. So you're
>> getting "2010" indexed as a separate token and
>> you're also searching on it...
>>
>> Best
>> Erick
>>
>> On Mon, Jun 13, 2011 at 3:07 PM, Tirthankar Chatterjee
>> <tchatter...@commvault.com> wrote:
>>> We are using edismax for query and the query fired is (url:_2010)
>>>
>>> http://redcarpet2.dm2.commvault.com:27000/solr/select/?q=url:
>>> 2010&version=2.2&start=0&rows=10&indent=on&defType=edismax<http://redcarpet2.dm2.commvault.com:27000/solr/select/?q=url:%202010&version=2.2&start=0&rows=10&indent=on&defType=edismax>
>>>
>>> the url field is of type text_rev
>>>
>>> Results that SOLR returns has 1 extra item which we don't want to get. How
>>> do we achieve that?
>>>
>>> Results:
>>>
>>> SPC265_SharePoint_2010.pptx
>>> OpenTRs2010.xlsx(we don't want this to be returned)
>>>
>>>
>>> Thanks in advance!!!
>>>
>>> Tirthankar
>>>
>>>
>>> ******************Legal Disclaimer***************************
>>> "This communication may contain confidential and privileged
>>> material for the sole use of the intended recipient. Any
>>> unauthorized review, use or distribution by others is strictly
>>> prohibited. If you have received the message in error, please
>>> advise the sender by reply email and delete the message. Thank
>>> you."
>>> *********************************************************
