Why is the constructor of TopFieldDocs not public?
I'm subclassing MultiSearcher and writing a customized searcher on my own. The search( Weight, Filter, int, Sort ) method on MultiSearcher should return TopFieldDocs, but I cannot instantiate one cause TopFieldDocs constructor is declared as package-private. (TopFieldDocs itself is public!) Is there any reason the constructor of TopFieldDocs has no modifier declaration? - Cheolgoo Kang - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
IndexWriter 2-phase commit usage
I'm having trouble applying IndexWriter 2-phase commit to make a transaction involving two different indexes. The scenario, 1. Open index1 2. Open index2 3. Make change1 to index1 4. Make change2 to index2 5. index1.prepareCommit() successfully flushes 6. index2.prepareCommit() successfully flushes 7. index1.commit() successfully (segmentN file written) 8. index2.commit() If step 8 fails, how can I roll back step 7 w/o implementing my own undo logic? I want to make change1 and change2 a single transaction. Thanks, An
Re: Upgrading an old index to 2.4 format
That is true. Good question on converting an already optimized index... I haven't tested this, and it's somewhat roundabout, but you may be able to open an IndexWriter, then call setUseCompoundFile with the opposite of what your index presently has, run optimize, set that back, and run optimize again. I think it should work. You could also add and then delete a single doc, then call optimize. Mike An Hong wrote: I've several existing indexes in 2.1 file format. If these indexes have multiple segments and have not been optimized, a call to IndexWriter.optimize() should create a newly merged segment of the latest 2.4 file format. Is this true? And, is there a way to force a file format upgrade on a 2.1 index that's already in optimized state? Thanks, An - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
Re: IndexWriter.rollback() logic question
The javadoc is not right -- I just committed a fix. rollback() discards all changes since the last commit, or since the IndexWriter was opened (if commit hasn't been called). So in your case, doc1 is in the index but doc2 is not. You can also call rollback() not having called prepareCommit(). Mike An Hong wrote: A question about IndexWriter.rollback() logic. Its javadoc says that it "Close the IndexWriter without committing any of the changes that have occurred since it was opened." So if I do 1. Open IndexWriter 2. Add doc1 3. Commit (successfully) 4. Add doc2 5. PrepareCommit() 6. Rollback() Is doc1 in the index? If not, I wonder how come rollback() does not roll back to the last successful commit() state instead? Thanks, An - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
Upgrading an old index to 2.4 format
I've several existing indexes in 2.1 file format. If these indexes have multiple segments and have not been optimized, a call to IndexWriter.optimize() should create a newly merged segment of the latest 2.4 file format. Is this true? And, is there a way to force a file format upgrade on a 2.1 index that's already in optimized state? Thanks, An
IndexWriter.rollback() logic question
A question about IndexWriter.rollback() logic. Its javadoc says that it "Close the IndexWriter without committing any of the changes that have occurred since it was opened." So if I do 1. Open IndexWriter 2. Add doc1 3. Commit (successfully) 4. Add doc2 5. PrepareCommit() 6. Rollback() Is doc1 in the index? If not, I wonder how come rollback() does not roll back to the last successful commit() state instead? Thanks, An
Re: Phrase indexing and searching with Lucene
: Subject: Phrase indexing and searching with Lucene : References: <499be497.2060...@mapmyindia.com> : <18248.30864...@web26005.mail.ukl.yahoo.com> http://people.apache.org/~hossman/#threadhijack Thread Hijacking on Mailing Lists When starting a new discussion on a mailing list, please do not reply to an existing message, instead start a fresh email. Even if you change the subject line of your email, other mail headers still track which thread you replied to and your question is "hidden" in that thread and gets less attention. It makes following discussions in the mailing list archives particularly difficult. See Also: http://en.wikipedia.org/wiki/Thread_hijacking -Hoss - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
Field Normalisation in Query across two indexes
Guys, I have a question on normalisation. I am using a prehistoric version of lucene; 2.0.0 Context: http://markmail.org/message/z5lcz2htjvqsscam I have these two scenario with indexes. One: 2 indexes; One with documents that has a field "field_one" instantiated as TOKENIZED and then setOmitNorms(true). Second index with the same field "field_one" instantiated TOKENIZED and then setBoost(boost_factor); Two: 2 indexes; One with field "field_one" instantiated as TOKENIZED and then setOmitNorms(true). Second index with the same field "field_one", but some of them are as first index and some were instantiated TOKENIZED and then setBoost(boost_factor); A mix of the two types. If I am to load these two indexes into an IndexReader and query on the field "field_one" would there be any issues that I should expect? Any advice is much appreciated. Cheerio
Re: Range search in dependent collections
Just repeat the first and last names in each document. How many users and jobs are you talking about? Lucene will cope unless the numbers are extremely large. Note that I'm not claiming this is necessarily the "best" way, just the simplest which is often, but not always, the best. -- Ian. On Mon, Feb 23, 2009 at 2:49 PM, Mykola Peleshchyshyn wrote: > What about the case with several jobs? I.e.: > > firstname: whatever > lastname: whatever > > jobtitle1: somejob > startdate1: whenever > enddate1: whenever > . > jobtitleN: somejob > startdateN: whenever > enddateN: whenever > > > > - Original Message > From: Ian Lea > To: java-user@lucene.apache.org > Sent: Monday, February 23, 2009 4:14:37 PM > Subject: Re: Range search in dependent collections > > Hi > > > The simplest way is to flatten your data into lucene documents that > hold all relevant info for user/job combinations. That will let you > search easily across whichever fields you want. > > Your documents could look like this: > > firstname: whatever > lastname: whatever > jobtitle: somejob > startdate: whenever > enddate: whenever > > and searches could be something like lastname:whatever jobtitle:somejob. > > > -- > Ian. > > > On Mon, Feb 23, 2009 at 9:12 AM, Mykola Peleshchyshyn > wrote: >> Hello, >> I'm new to Lucene and I have the following problem: >> >> I have a Users with first name, last name etc. and User Jobs (collection) >> with job title, start date, end date. >> >> I need to perform search on all user fields (fist name, last name etc) and >> job fields (job title and optionally start date and end date). >> >> What would be the best way to index and search this data? >> >> Thank you, >> Mykola Peleshchyshyn - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
Re: Range search in dependent collections
What about the case with several jobs? I.e.: firstname: whatever lastname: whatever jobtitle1: somejob startdate1: whenever enddate1: whenever . jobtitleN: somejob startdateN: whenever enddateN: whenever - Original Message From: Ian Lea To: java-user@lucene.apache.org Sent: Monday, February 23, 2009 4:14:37 PM Subject: Re: Range search in dependent collections Hi The simplest way is to flatten your data into lucene documents that hold all relevant info for user/job combinations. That will let you search easily across whichever fields you want. Your documents could look like this: firstname: whatever lastname: whatever jobtitle: somejob startdate: whenever enddate: whenever and searches could be something like lastname:whatever jobtitle:somejob. -- Ian. On Mon, Feb 23, 2009 at 9:12 AM, Mykola Peleshchyshyn wrote: > Hello, > I'm new to Lucene and I have the following problem: > > I have a Users with first name, last name etc. and User Jobs (collection) > with job title, start date, end date. > > I need to perform search on all user fields (fist name, last name etc) and > job fields (job title and optionally start date and end date). > > What would be the best way to index and search this data? > > Thank you, > Mykola Peleshchyshyn - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
Re: Range search in dependent collections
Hi The simplest way is to flatten your data into lucene documents that hold all relevant info for user/job combinations. That will let you search easily across whichever fields you want. Your documents could look like this: firstname: whatever lastname: whatever jobtitle: somejob startdate: whenever enddate: whenever and searches could be something like lastname:whatever jobtitle:somejob. -- Ian. On Mon, Feb 23, 2009 at 9:12 AM, Mykola Peleshchyshyn wrote: > Hello, > I'm new to Lucene and I have the following problem: > > I have a Users with first name, last name etc. and User Jobs (collection) > with job title, start date, end date. > > I need to perform search on all user fields (fist name, last name etc) and > job fields (job title and optionally start date and end date). > > What would be the best way to index and search this data? > > Thank you, > Mykola Peleshchyshyn - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
Range search in dependent collections
Hello, I'm new to Lucene and I have the following problem: I have a Users with first name, last name etc. and User Jobs (collection) with job title, start date, end date. I need to perform search on all user fields (fist name, last name etc) and job fields (job title and optionally start date and end date). What would be the best way to index and search this data? Thank you, Mykola Peleshchyshyn - To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
problem with indexing
hi all, I want to modify the indexing process lucene to suit my application. I partially read the book LUCENE IN ACTION but it contains no information to extract the term document matrix. I want to apply LSI to the indexed document. Please somebody help me out. regards nitin