lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Upayavira ...@odoko.co.uk>
Subject Re: How to perform phonetic matching/query for multivalued string field
Date Wed, 16 Sep 2015 22:39:18 GMT
I bet the terms component does not analyse the terms, so you will need
to hand in already analysed phonetic terms. You could use the
http://localhost:8983/solr/YOUR-CORE/analysis/field URL to have Solr
analyse the field for you before passing it back to the term component.

Upayavira

On Wed, Sep 16, 2015, at 10:03 PM, Jie Gao wrote:
> Many thanks for your suggestion.
> 
> It works well for querying the field with phonetic matching and responses
> a
> list of docs tagged with the term.
> 
> However, is there any way that i can get a list of matched terms ? The
> phonetic matching seems not work with Term Component (i'm using
> terms.regex
> to filter).
> 
> Jie Gao,
> Research Assistant,
> Department of Computer Science, The University of Sheffield,
> Regent Court, 211 Portobello, S1 4DP, Sheffield, UK
> 
> On 16 September 2015 at 21:04, Upayavira <uv@odoko.co.uk> wrote:
> 
> > That is, use a TextField plus a KeywordTokenizerFactory, rather than a
> > StringField
> >
> > On Wed, Sep 16, 2015, at 09:03 PM, Upayavira wrote:
> > > If you want to analyse a string field, use the KeywordTokenizer - it
> > > just passes the whole field through as a single tokenizer.
> > >
> > > Does that get you there?
> > >
> > > On Wed, Sep 16, 2015, at 08:52 PM, Jie Gao wrote:
> > > > I understand that i can configure "solr.PhoneticFilterFactory" for both
> > > > indexing and query time for "solr.TextField". However, i want to query
> > a
> > > > list of term (indexed and stored) from a field ordered by phonetic
> > > > similarity, which can be easily done by most of relational database.
> > > >
> > > > Term Component allows me to perform exactly matching and regex based
> > > > fuzzy
> > > > matching from multi-valued field. However, the solr string field does
> > not
> > > > allow to customise the default analyser. Is there any other way to
> > > > circumvent the problem?
> > > >
> > > > thanks,
> > > > Jerry
> > > >
> > > >
> > > >
> > > > On 16 September 2015 at 19:55, Upayavira <uv@odoko.co.uk> wrote:
> > > >
> > > > >
> > > > >
> > > > > On Wed, Sep 16, 2015, at 06:37 PM, Jie Gao wrote:
> > > > > > Hi,
> > > > > >
> > > > > >
> > > > > > I want to query a list of terms indexed and stored in multivalued
> > string
> > > > > > field via Term Component. The term component can support exact
> > matching
> > > > > > and
> > > > > > regex based fuzzy matching. However, Is any way i can configure
> > scheme to
> > > > > > do phonetic matching/query?
> > > > >
> > > > > Phonetic matching is done at index time - that is - you use a
> > > > > PhoneticFilterFactory in your analysis chain, such that you are doing
> > > > > exact match lookups on the phonetic terms.
> > > > >
> > > > > Make sense?
> > > > >
> > > > > Upayavira
> > > > >
> >

Mime
View raw message