Applying analyzer in wildcard queries

Rike · **Joined:** Fri Apr 20, 2012 7:07 am **Posts:** 2

Hi,
we are trying to use HibernateSearch in our application because we need the possibility of fuzzy searching. So far we basically use SQL queries on the database.
Now we have the problem of how to use both fuzzy and wildcard queries on the same index.
It is recommended to use the same analyzer on both indexing and searching. So we defined an analyzer that uses StandardFilter, LowerCaseFilter and ASCIIFoldingFilter.
But wildcard queries do not apply the analyzer on the matching terms, so e.g. "Äpfel" doesn't find "Äpfel und Birnen".
These two rules ("use the same analyzer" and "wildcard queries do not apply the analyzer") seem to be conflicting - or do we miss anything here?
Is there any way to force a wildcard query to apply the analyzer? Or to apply the analyzer's filters on our search term before building the wildcard query?

Thanks for any hint,
Friederike

sanne.grinovero · **Posted:** Mon Apr 23, 2012 6:12 am

Hi Friederike,
do you have a suggestion on how we could improve the QueryBuilder ? Maybe if you could provide a working testcase we can make this easier.

In the meantime you could indeed filter the query terms yourself by using the utilities in

Code:

org.hibernate.search.util.AnalyzerUtils

Rike · **Joined:** Fri Apr 20, 2012 7:07 am **Posts:** 2

Hi Sanne,

thanks for the suggestion. Unfortunately we are bound to Hibernate 3.6.6, ergo to HibernateSearch 3.4.1 which doesn't yet provide AnalyzerUtils.
I guess I have to write my own normalizing procedure until we can move to the newest version of HibernateSearch.

When we tried to solve our problem we hoped to find a way to force wildcard queries to apply an analyzer. But ConnectedMultiFieldsTermQueryBuilder.getAllTermsFromText() generally doesn't use the Analyzer if Approximation is WILDCARD. This is confusing since createQuery() only calls getAllTermsFromText if the FieldContext is not set to ignore the analyzer.

It would be great having a way to tell the QueryBuilder to use the analyzer independent from the TermContext.Approximation.
Or, even better, to have the QueryBuilder respect the FieldContext's ignoreAnalyzer flag and set the flag to true when creating a WildCardContext.

I still don't get why wildcard queries do ignore the analyzer at all. Following the rule of thumb to use the same analyzer for indexing and searching I should be able to use the analyzer for all kinds of queries.
What am I missing here?

Thanks for your help,
Friederike