FAQ
Hello everyone,


I noticed that when updating a large index with IndexModifier (lots
of delete+add), the bottlenek was the delete. If I only add, speed is
at least multiplied by 50.
From what i have understood, its because a new IndexReader is
initialized for each delete.

Is the new updateDocument() method from IndexWriter optimized for
this kind of problem?


Thanks!


--
Antoine Baudoux
Development Manager
ab@taktik.be
Tél.: +32 2 333 58 44
GSM: +32 499 534 538
Fax.: +32 2 648 16 53



---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org

Search Discussions

  • Mark Miller at Sep 22, 2007 at 9:30 pm
    Yes. Its much faster. The deletes are buffered and an IndexReader is
    only opened to flush a specified number of buffered deletes.
    IndexModifier is terribly inefficient and I think, to be deprecated soon.

    - Mark

    Antoine Baudoux wrote:
    Hello everyone,


    I noticed that when updating a large index with IndexModifier
    (lots of delete+add), the bottlenek was the delete. If I only add,
    speed is at least multiplied by 50.
    From what i have understood, its because a new IndexReader is
    initialized for each delete.

    Is the new updateDocument() method from IndexWriter optimized for
    this kind of problem?


    Thanks!


    --
    Antoine Baudoux
    Development Manager
    ab@taktik.be
    Tél.: +32 2 333 58 44
    GSM: +32 499 534 538
    Fax.: +32 2 648 16 53



    ---------------------------------------------------------------------
    To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
    For additional commands, e-mail: java-user-help@lucene.apache.org
    ---------------------------------------------------------------------
    To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
    For additional commands, e-mail: java-user-help@lucene.apache.org
  • Tom Conlon at Sep 23, 2007 at 8:55 am
    Hi,

    Unless I'm missing something, according to:

    http://lucene.apache.org/java/docs/queryparsersyntax.html#Escaping%20Spe
    cial%20Characters

    I should be able to search for C++ and C# using something like:
    C\+\+ and C\#.

    This doesn't work.

    I apologise if this sort of question has been asked many times in the
    past but I've not found any examples or answer to this problem.

    Thanks,
    Tom

    ---------------------------------------------------------------------
    To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
    For additional commands, e-mail: java-user-help@lucene.apache.org
  • Karl Wettin at Sep 23, 2007 at 9:47 am
    23 sep 2007 kl. 10.53 skrev Tom Conlon:
    Unless I'm missing something, according to:

    http://lucene.apache.org/java/docs/queryparsersyntax.html#Escaping%
    20Spe
    cial%20Characters

    I should be able to search for C++ and C# using something like:
    C\+\+ and C\#.
    That is correct.
    This doesn't work.
    Did you use the same analyzer when populating the index as when you
    create the query?

    If you did, can you demonstrate the problem with a small stand alone
    test case?


    --
    karl

    ---------------------------------------------------------------------
    To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
    For additional commands, e-mail: java-user-help@lucene.apache.org
  • Tom Conlon at Sep 23, 2007 at 11:18 am
    Hi Karl,
    Did you use the same analyzer when populating the index as when you
    create the query?

    Yes, I used the default analyzer in a modified version of the command
    line demo.
    If you did, can you demonstrate the problem with a small stand alone
    test case?

    I'll try.

    Tom

    -----Original Message-----
    From: Karl Wettin
    Sent: 23 September 2007 10:39
    To: java-user@lucene.apache.org
    Subject: Re: Escaping special characters


    23 sep 2007 kl. 10.53 skrev Tom Conlon:
    Unless I'm missing something, according to:

    http://lucene.apache.org/java/docs/queryparsersyntax.html#Escaping%
    20Spe
    cial%20Characters

    I should be able to search for C++ and C# using something like:
    C\+\+ and C\#.
    That is correct.
    This doesn't work.
    Did you use the same analyzer when populating the index as when you
    create the query?

    If you did, can you demonstrate the problem with a small stand alone
    test case?


    --
    karl

    ---------------------------------------------------------------------
    To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
    For additional commands, e-mail: java-user-help@lucene.apache.org


    ---------------------------------------------------------------------
    To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
    For additional commands, e-mail: java-user-help@lucene.apache.org

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupjava-user @
categorieslucene
postedSep 22, '07 at 9:24p
activeSep 23, '07 at 11:18a
posts5
users4
websitelucene.apache.org

People

Translate

site design / logo © 2022 Grokbase