Lucene Query

[pascal]

Hi,

I'm currently working at a C# Project using web services.
I created 2 custom attributes:

- language (text)
- type  (text)

I'd like to search all the documents that are from a specified document type.

This code snippet returns me all the documents that are from type "KT-HT":

// Create a query object, looking for all items with alfresco in the name of text

RepositoryService.Query query = new RepositoryService.Query();
query.language = RepositoryService.QueryLanguageEnum.lucene;

query.statement = "@my\\:type:'KT-HT'";

 // Execute the query
 QueryResult queryResult = this.repoService.query(this.spacesStore, query, false);‍‍‍‍‍‍‍‍‍

       
But if I want tu use wildcards in the query statement like this:

query.statement = "@my\\:type:'KT*'";‍

I get an empty resultset.

Any clue of what I am doing wrong would be greatly appreciated.

Thanks a lot.

Pascal

[andy]

Hi

query.statement = "@my\\:type:KT*";
‍‍‍

Should work.

If the text is in double quotes it is treated as a phrase (no wild cards allowed). See http://lucene.apache.org/java/docs/queryparsersyntax.html

I am a bit confused here as phrases should be surrounded by double quotes and you have single quotes.

Your text will be tokenised when indexed. Depending on the attribute type and config this probably goes in as two tokens KT and HT. KT* should work.


You may also consider:
* grouping things using categories
* having real types and using the TYPE:"….." lucene attribute (which will expand to include all subtypes of the specified type)


Regards

Andy

[pascal]

Hi Andy,

Thank you very much for your quick answer.

I tried with

query.statement = "@my\\:type:KT*"; ‍

But I still get an empty resultset.

Here is the defintion of my custom attribute:

<types>
       <type name="my:customeType">
         <title>Type</title>
         <parent>cm:content</parent>
         <properties>
            <property name="my:type>
                <type>d:text</type>
            </property>
         </properties>
</type>

Did I miss something in the definition or in the config files ?

Thank you for your help.

Pascal

[andy]

Hi

The model looks OK.

I forgot that the text tokeniser will be converting everything to lower case.
So you will need  

query.statement = "@my\\:type:kt*";

This form of query (wild card) does not go through the tokeniser so will be used "as is".

If you want an exact match you will need

query.statement = "+@my\\:type:kt  +@my\\:type:ht";

or

query.statement = "+@my\\:type:\"KT-HT\"";


Apologies

Regards

Andy

[pascal]

Hi Andy,

I finally managed to get a correct resultset by using:

query.statement = "+@\\{my.new.model\\}type:kt*) ";‍

It was the only way for me to use wildcards in the query.

Thank your for your help.
I am very satisfied with your amazing product.

Regards.

Pascal