Hibernate Search: need exact match on a specific field

liudan2005 · **Joined:** Fri Dec 22, 2006 8:17 am **Posts:** 6

I need to make following HSQL to a lucene queue in hibernate search:
"from Person where name like '%james%' and country='us' "

I've tried:
luceneQuery = parser.parse( "name:james And country:us" )
this doesn't work because country "usa" is returned.

How to make this query? Anyone knows?

emmanuel · **Posted:** Wed Apr 18, 2007 7:03 pm

interesting, did you tokenize the field? And what is your analyzer?

emmanuel · **Posted:** Wed Apr 18, 2007 7:04 pm

Also, doing lucenequery.toString() helps to understand what happens

liudan2005 · **Joined:** Fri Dec 22, 2006 8:17 am **Posts:** 6

I'm using ChineseAnalyzer. Is it possible to use ChineseAnalyzer for name field, and use KeywordAnalyzer for country field in my query? I think the problem can be solved if I can use different analyzers for those fields.

emmanuel · **Posted:** Fri Apr 20, 2007 10:40 am

no, not from the regular Lucene query parser.
But you can:
- write your own query parser by extending theirs
- use the programmatic API Lucene Query objects, to tune what you want (which is usually what people end up doing with complex query requirements.

mirak · **Joined:** Mon Apr 04, 2011 12:08 pm **Posts:** 32

emmanuel wrote:

- use the programmatic API Lucene Query objects, to tune what you want (which is usually what people end up doing with complex query requirements.

what would you use to have an exact match with a query ?

.ignoreAnalyzer() ?

emmanuel · **Posted:** Fri Apr 22, 2011 6:13 am

With the Hibernate Search query DSL, you would do

Code:

builder.keyword().onField("name").matching("Emmanuel").createQuery()

mirak · **Joined:** Mon Apr 04, 2011 12:08 pm **Posts:** 32

the initial problem seems that 'emmanuell' would match 'emmanuel'

sanne.grinovero · **Posted:** Fri Apr 22, 2011 7:42 am

Hi mirak,
use

Code:

@Field(index=Index.UN_TOKENIZED

on the property that you don't want to analyze.

Or alternatively specify a different analyzer to it which fits your need: apparently you don't want to use the ChineseAnalyzer on that property, but you can apply different analyzers to each field. (UN_TOKENIZED means no analyzer at all is going to be applied, treating the full value as a keyword)

mirak · **Joined:** Mon Apr 04, 2011 12:08 pm **Posts:** 32

ok, thank you