Hi everybody. I'm using hibernate-search 3.0.0.GA and found some strange things.

By example, if i run search with query, which contatins special characters (/, \, {, }, * and so on), and i have records with these characters in my DB , then i get empty search result. Why? Before search i call "QueryParser.escape(query)", so all special chars must be escaped after this.

By the way, the * character is some kind of magical sign. If i want to search records, which contains *** (three asterisks)(by example: fu***ng, pet***g and so on), i must call queryParser.parse with following parameter: *\*\*\** , right? 3 escaped asterisks and leading and ending asterisks for wildcarding. But this query causes some kind of "broadcast" query and i have "BooleanQuery:TooManyClauses" exception. Why? What should i do to search these kind of words? By the way, if i enter *** in Hibernate Forum search, i think, it crashes, because i waited about 5 minutes for search results, but nothing happens. :)

And another one thing. If i run search with word "and", search engine just ignore it. Is that some special word? How can i disable it special meaning?

Hi,

I might not have the answers for all your questions, but maybe a few. Generally I have the impression that your problems are more Lucene related than Hibernate Search. I recommend you to read some of the Lucene documentation.

First of all this http://lucene.apache.org/java/docs/queryparsersyntax.html. For example, the boolean and operator is AND (all CAPS) or &&. And yes, the backslash escapes the wildcard functionality of *. However, if you start a query with a wildcard star as in your example it is quite easy to get a TooManyClausesEsception. This topic and some workarounds are explained here http://wiki.apache.org/lucene-java/LuceneFAQ#head-06fafb5d19e786a50fb3dfb8821a6af9f37aa831.

Last but not least - check your Lucene index via Luke to see if the special characters got indexed the way you expected (http://hibernate.org/440.html). Some analyzers might remove these special characters when indexing. Are you indexing your data tokenized or un-tokenized?

--Hardy

Thanks a lot for your reply. I will try what you say.
PS. I'm indexing my data with "tokenized" flag.