Mailing-List: contact solr-user-help@lucene.apache.org; run by ezmlm
Precedence: bulk
Reply-To: solr-user@lucene.apache.org
Received-SPF: pass (nike.apache.org: domain of tech178@yoox.com designates
 195.130.217.15 as permitted sender)
From: Finotti Simone <tech178@yoox.com>
To: "solr-user@lucene.apache.org" <solr-user@lucene.apache.org>
Subject: Re: Skip first word
Thread-Topic: Skip first word
Thread-Index: 
 Ac1qfxjURy/VT/HYRD+Ya0bgkqD9Zf//4EEAgAEYHx2AAIBWAIABNbpn///kCQCAACi3hw==
Date: Fri, 27 Jul 2012 09:46:47 +0000
Message-ID: <F7D5CF98F585454EBFC370E8D0268C8E3484461A@Ymail13blq.yoox.net>
References: 
 <F7D5CF98F585454EBFC370E8D0268C8E3483F1AF@Ymail13blq.yoox.net>,<1343232620.20860.YahooMailClassic@web121703.mail.ne1.yahoo.com>
 <F7D5CF98F585454EBFC370E8D0268C8E34842330@Ymail13blq.yoox.net>,<E4F0E68C-3CA1-4E77-B113-4F2F94320444@it-agenten.com>
 <F7D5CF98F585454EBFC370E8D0268C8E348444AD@Ymail13blq.yoox.net>,<01ADFE28-570B-4AAD-85A1-C499B863986A@it-agenten.com>
In-Reply-To: <01ADFE28-570B-4AAD-85A1-C499B863986A@it-agenten.com>
Accept-Language: it-IT, en-US
Content-Language: it-IT
MIME-Version: 1.0
Content-Type: text/plain; charset=WINDOWS-1252
Content-Transfer-Encoding: quoted-printable

Brilliant!=0AThank you very much :)=0A=0A__________________________________=
______=0AInizio: Chantal Ackermann [c.ackermann@it-agenten.com]=0AInviato: =
venerd=EC 27 luglio 2012 11.20=0AFine: solr-user@lucene.apache.org=0AOggett=
o: Re: Skip first word=0A=0AHi Simone,=0A=0Ano I meant that you populate th=
e two fields with the same input - best done via copyField directive.=0A=0A=
The first field will contain ngrams of size 1 and 2. The other field will c=
ontain ngrams of size 3 and longer (you might want to set a decent maxsize =
there).=0A=0AThe query for the autocomplete list uses the first field when =
the input (typed in by the user) is one or two characters long. Your exampl=
e was: "D", "G", or than "Do" or "Ga". The result would search only on the =
single token field that contains for the input "Dolce & Gabbana" only the n=
grams "D" and "Do". So, only the input "D" or "Do" would result in a hit on=
 "Dolce & Gabbana".=0AOnce the user has typed in the third letter: "Dol" or=
 "Gab", you query the second, more tokenized field which would contain for =
"Dolce & Gabbana" the ngrams "Dol" "Dolc" "Dolce" "Gab" "Gabb" "Gabba" etc.=
=0ABoth inputs "Gab" and "Dol" would then return "Dolce & Gabbana".=0A=0A1.=
 First  field type:=0A=0A<tokenizer class=3D"solr.KeywordTokenizerFactory"/=
>=0A<filter class=3D"solr.EdgeNGramFilterFactory" minGramSize=3D"1" maxGram=
Size=3D"2" side=3D"front"/>=0A=0A2. Secong field type:=0A=0A<tokenizer clas=
s=3D"solr.WhitespaceTokenizerFactory"/>=0A<!-- maybe add WordDelimiter etc.=
 -->=0A<filter class=3D"solr.EdgeNGramFilterFactory" minGramSize=3D"3" maxG=
ramSize=3D"10" side=3D"front"/>=0A=0A3. field declarations:=0A=0A<field nam=
e=3D"short_prefix" type=3D"short_ngram" =85 />=0A<field name=3D"long_prefix=
" type=3D"long_ngram" =85 />=0A=0A<copyField source=3D"short_prefix" dest=
=3D"long_prefix" />=0A=0A=0AChantal=0A=0AAm 27.07.2012 um 11:05 schrieb Fin=
otti Simone:=0A=0A> Hi Chantal,=0A>=0A> if I understand correctly, this imp=
lies that I have to populate different fields according to their lenght. Si=
nce I'm not aware of any logical condition you can apply to copyField direc=
tive, it means that this logic has to be implementend by the process that p=
opulates the Solr core. Is this assumption correct?=0A>=0A> That's kind of =
bad, because I'd like to have this kind of "rules" in the Solr configuratio=
n. Of course, if that's the only way... :)=0A>=0A> Thank you=0A>=0A> ______=
__________________________________=0A> Inizio: Chantal Ackermann [c.ackerma=
nn@it-agenten.com]=0A> Inviato: gioved=EC 26 luglio 2012 18.32=0A> Fine: so=
lr-user@lucene.apache.org=0A> Oggetto: Re: Skip first word=0A>=0A> Hi,=0A>=
=0A> use two fields:=0A> 1. KeywordTokenizer (=3D single token) with ngram =
minsize=3D1 and maxsize=3D2 for inputs of length < 3,=0A> 2. the other one =
tokenized as appropriate with minsize=3D3 and longer for all longer inputs=
=0A>=0A>=0A> Cheers,=0A> Chantal=0A>=0A>=0A> Am 26.07.2012 um 09:05 schrieb=
 Finotti Simone:=0A>=0A>> Hi Ahmet,=0A>> business asked me to apply EdgeNGr=
am with minGramSize=3D1 on the first term and with minGramSize=3D3 on the l=
atter terms.=0A>>=0A>> We are developing a search suggestion mechanism, the=
 idea is that if the user types "D", the engine should suggest "Dolce & Gab=
bana", but if we type "G", it should suggest other brands. Only if users ty=
pe "Gab" it should suggest "Dolce & Gabbana".=0A>>=0A>> Thanks=0A>> S=0A>> =
________________________________________=0A>> Inizio: Ahmet Arslan [iorixxx=
@yahoo.com]=0A>> Inviato: mercoled=EC 25 luglio 2012 18.10=0A>> Fine: solr-=
user@lucene.apache.org=0A>> Oggetto: Re: Skip first word=0A>>=0A>>> is ther=
e a tokenizer and/or a combination of filter to=0A>>> remove the first term=
 from a field?=0A>>>=0A>>> For example:=0A>>> The quick brown fox=0A>>>=0A>=
>> should be tokenized as:=0A>>> quick=0A>>> brown=0A>>> fox=0A>>=0A>> Ther=
e is no such filter that i know of. Though, you can implement one with modi=
fying source code of LengthFilterFactory or StopFilterFactory. They both re=
move tokens. Out of curiosity, what is the use case for this?=0A>>=0A>>=0A>=
>=0A>>=0A>=0A>=0A>=0A>=0A>=0A=0A=0A=0A