ScopedAnalyzer contains PassThroughAnalyzer on Class Level

timomeinen · **Joined:** Wed May 31, 2006 2:34 pm **Posts:** 9

Hi,

I am trying to figure out, why my QueryParser doesn't parse with the correct analyzer. I noticed, that ScopedAnalyzer is used, to handle the different analyzers, which are field dependent.

The problem is, that the ScopedAnalyzer contains only PassThroughAnalyzers, when the analyzer of the entity is defined at class level with @Analyzer. Only the fields marked with @Field(analyzer = @Analyzer(...)) are represented by a DelegateNamedAnalyzer.

Is it a bug or do I use the api wrong?

This is the way, I get the analyzer of a class (Profile in this example):

Code:

Search.getFullTextEntityManager(this.entityManager).getSearchFactory().getAnalyzer(Profile.class);

BTW: Is there a way available to get an anylzer instance, which is defined by an @AnalyzerDef annotation? I do all of the above, just to get an anylzer with a special synonymfilter for the QueryParser.

Thank you.
Timo

hardy.ferentschik · **Posted:** Tue Nov 09, 2010 9:44 am

Hi Timo,

the ScopedAnalyzer should take care transparently which analyzer to chose (also at query time). If the field name you are trying to index does not specify a more specific analyzer the global analyzer (which is the analyzer you defined on class level) is used. I don't think that the ScopedAnalyzer should only contain PassThroughAnalyzers.
Maybe you can put your question into some context. What usecase do you try to solve? How does your configuration look like and your query building code?

--Hardy

sanne.grinovero · **Posted:** Mon Nov 15, 2010 6:06 pm

Quote:

BTW: Is there a way available to get an anylzer instance, which is defined by an @AnalyzerDef annotation? I do all of the above, just to get an anylzer with a special synonymfilter for the QueryParser.

It seem you missed the easy way ;)

timomeinen · **Joined:** Wed May 31, 2006 2:34 pm **Posts:** 9

Hi,

yes, the easy way is sometimes the best way. In the end I am using now the getAnalyzer() method to get the analyzer and use it to create a TokenStream:

Code:

TokenStream stream = analyzer.tokenStream(field, reader);

Anyhow, originally I wanted QueryParser to automatically select the right analyzer for the specified field, which didn't work, because the ScopedAnalyzer only contains instances of PassThroughAnalyzer, when the @Analyzer is defined on class/type level.

When fields are annotated with @Analyzer, then instances of DelegateNameAnalyzer are within the ScopedAnalyzer and the QueryParser uses the correct analyzer. This seems to me like a bug, because I would expect, that an class level analyzer is used on every field in its class.

sanne.grinovero · **Posted:** Tue Nov 16, 2010 11:02 am

yes it might be a bug: please open a JIRA issue and attach a testcase. Of course fixes are welcome too.