Hi Alex,
It seems that you're using a stemmer in your analyzer chain when indexing
documents. So the term "consciousness" will be stemmed and stored as
"conscious" in the database. This means that you have to process your query
with the same analyzer that's been used for indexing. Instead of
$query = Lucy::Search::TermQuery->new(
field => $field,
term => $token,
);
you should use something like the following:
my $type = $schema->fetch_type($field);
my $analyzer = $type->get_analyzer;
my $terms = $analyzer->split($token);
if ( @$terms == 1 ) {
$query = Lucy::Search::TermQuery->new(
field => $field,
term => $terms->[0],
);
}
elsif ( @$terms > 1 ) {
$query = Lucy::Search::PhraseQuery->new(
field => $field,
terms => $terms,
);
}
Lucy's built-in QueryParser class will do that for you. If you want to use a
custom query parser, you can find some documentation here:
http://lucy.apache.org/docs/perl/Lucy/Docs/Cookbook/CustomQueryParser.html
Nick
On Dec 25, 2012, at 15:03 , Aleksandar Radovanovic <[email protected]>
wrote:
> Hi there,
>
> I was trying to match a term exactly by using
>
> Lucy::Search::PhraseQuery and Lucy::Search::TermQuery
>
> but I could not get the exactly-exact match. For example, if I try to
> find the term "consciousness" (my text is full of it :-) I do not have
> any matches. However, the search for "conscious" gives me matches to
> "conscious" and to "consciousness". Since those are two different
> concepts I consider the result being incorrect.
>
> Is there any way to get the exactly-exact match? (or perhaps I am doing
> something wrong).
>
> Alex
