lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ronan KERDUDOU (JIRA)" <>
Subject [jira] Commented: (LUCENE-1100) StandardTokenizer incorrectly types certain values
Date Mon, 01 Mar 2010 13:20:07 GMT


Ronan KERDUDOU commented on LUCENE-1100:

when having in a text : "5.3.2-blàhblah" this is detected as a NUM... and "2.45" is detected
as a HOST...
I think when there's a mix of alpha and num the '-' should be considered as a whitespace...
and that a {NUMERIC}[.,]{NUMERIC} must be a NUM and not a HOST.
also something like "image.gif" shouldn't be a HOST...
I know it's very difficult to make this flex file perfect.

"TEST" and "2010" are ALPHANUM, I would prefer to distinguish a ALPHA and a NUMERIC instead
"2,5" and "9-juin-2009" are NUM, i would prefer to distinguish NUM_DECIMAL and NUM_LIKE

i think sometimes we would have a better index if we could have 2 tokens with the same part
of the file.
for exemple : "5.3.2-blàhblah" would make a token but another token "blàhblah" should be
maybe we should do it in the StandardTokenizer to complete the power of the flex file ?

> StandardTokenizer incorrectly types certain values
> --------------------------------------------------
>                 Key: LUCENE-1100
>                 URL:
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Analysis
>    Affects Versions: 2.2
>            Reporter: Grant Ingersoll
>            Priority: Minor
> The StandardTokenizer incorrectly marks floating point values as <HOST>.  See
 There may be other things like this that are marked as incorrectly as well.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message