Thanks Digy, but if I index files i.e. xls, logs instead of database then how to find distinct from Index ?
On 5/30/07, Digy <[EMAIL PROTECTED]> wrote:
Hi Laxmilal Why don't you use a database for your application. It seems like a DB application. DIGY -----Original Message----- From: Laxmilal Menaria [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 30, 2007 11:52 AM To: [email protected]; [EMAIL PROTECTED] Subject: Re: Get all unique values of specific field Thanks, Thats okay for short index, But if index have millions of records or GB's data then it will get slow . So what is better ? On 5/30/07, Erich Eichinger <[EMAIL PROTECTED]> wrote: > > > hi, > > I guess it doesn't work without any iteration. > > Another option coming to my mind: If the list of possible classnames isn't > too long, you could do (pseudocode) > > Hashtable resultCount = new Hashtable(); > foreach( string classname in possibleClassNames ) > { > resultlist = index.SearchFor("Name:Menaria AND Class:"+classname) > resultCount[classname] = resultlist.Count; > } > > cheers, > Erich > > > > -----Original Message----- > > From: Michael Mitiaguin [mailto:[EMAIL PROTECTED] > > Sent: Wednesday, May 30, 2007 9:33 AM > > To: [email protected] > > Subject: Re: Get all unique values of specific field > > > > Laxmilal , > > > > What's the problem with a hit list iteration ( it should be > > very fast ) ? > > I am not sure about equivalent of SQL "distinct" in Lucene. > > You didn't describe whether you index ( and plus store ) all fields. > > Effectively you may just store a primary key ( let's say > > incremental int id for the sake of example , though you > > table doesn't look normalised , it doesn't really matter in > > our case ) which will be stored not indexed , the rest > > may be indexed as ( Name + Address + Class ...) or as > > separate fields if you need to and either stored or not stored. > > Having applied search "Name:Menaria" ( and having written > > all this I realised that you still need to iterate a hit list > > but let me finish > > :) ) you may form a string with coma delimited IDs and then > > > > "select distinct class from mytable where where id in "(" + > > formedstring + ") " > > > > Surely if you store everything in Lucene index , there is no > > need to query database and you may use collections for > > picking up distinct "Class" values , but my understanding > > you still need to iterate through Hits > > > > Regards > > Michael > > > > On 5/30/07, Laxmilal Menaria <[EMAIL PROTECTED]> wrote: > > > Hello everyone, > > > > > > I have created a Lucene Index of Students Database, this > > database have > > > 5 fields i.e. Name, Address, Class, PhoneNo and ScholarNo. > > > Now I have opened Searcher and query "Name:Menaria" , Its > > return 100 > > > results. Now I wants the All unique "Class" names which is > > return in > > > Hits objects, How can I get unique Class list without using Loop. > > > > > > Please suggest me.. > > > > > > -- > > > Thanks, > > > Laxmilal menaria > > > > > > http://www.minalyzer.com/ > > > http://www.chambal.com/ > > > > > > -- Thanks, Laxmilal menaria http://www.minalyzer.com/ http://www.chambal.com/
-- Thanks, Laxmilal menaria http://www.minalyzer.com/ http://www.chambal.com/
