lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erick Erickson" <erickerick...@gmail.com>
Subject Re: RE : blank space before special characters
Date Mon, 05 Nov 2007 18:34:59 GMT
There are several issues here....
1> How are you getting the entity reference? You must be encoding
the stream (or getting it encoded for you). So the first thing I'd do
is un-encode it.
2> After that, it's a question of what Filters/Analyzers you're using.
Take a look at ISOLatin1AccentFilter. I'm unclear whether it "closes up" the
case you're looking at, so be sure to check.
3> Since my peculiar situation can't use the Filter (the character
set I'm using isn't standard), I've pre-processed the input (both at
index and query time) to substitute the empty string for the
apostrophe

Hope this helps
Erick

On 11/5/07, Leire Urcelay <Leire.Urcelay@unil.ch> wrote:
>
> Sorry, I did a mistake in my previous email.
> The field "L'article" is indexed as "L &apos;article". The blank space is
> inserted between 'L' and '&apos;article'.
>
> Thanks,
>
> Leire
>
> -----Message d'origine-----
> De: Leire Urcelay [mailto:Leire.Urcelay@unil.ch]
> Envoyé: lundi, 5. novembre 2007 13:02
> À: java-user@lucene.apache.org
> Objet: blank space before special characters
>
> Hello,
>
> I have the following problem with my lucene index.
>
> When indexing fields containing special characters (like &), a blank space
> is inserted before the special character. For example: the content
> "L'article" is indexed as "L &apos;" (with a blank space between 'L' and
> '&amp;').
>
> Is there any way to avoid that?
>
> The characteristics of my field are the following: Indexed, Tokenized,
> Stored and Term Vector.
>
> Thanks in advance for your help,
>
> Leire
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message