lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From John Smith <john_smith9...@yahoo.com>
Subject ParallelReader and Date Filter
Date Mon, 12 Sep 2005 18:04:41 GMT
Hi, 
 
I have Lucene 1.4.3 codebase and I got Parallel Reader from the trunk along with a few changes
that need to go on top of it to make it compile.
 
II have 2 indexes, against which I am querying using the Parallel Reader. Most of my queries
work great. Thanks for the great work on this feature.
 
One index has only content and the other index has dates and basically all the information
related to the document.
 
 I have two questions.  
Question 1: 

========
When I use a date filter (using the DateFilter.java) in my query, then it seems to hang. Has
anyone faced an issue with using DateFilter  and Parallel Reader. 
 
On further debugging, I found it seems to continously go through this piece of code in FSDirectory:
FSInputStream.clone method continously.
  public Object clone() {
      FSInputStream clone = (FSInputStream)super.clone();
      clone.isClone = true;
    return clone;
  }

This seems to be the code path taken...

        at org.apache.lucene.store.FSInputStream.clone(FSDirectory.java)
        at org.apache.lucene.index.SegmentTermEnum.clone(SegmentTermEnum.java)
        at org.apache.lucene.index.TermInfosReader.terms(TermInfosReader.java)
        at org.apache.lucene.index.SegmentReader.terms(SegmentReader.java)
        at org.apache.lucene.index.ParallelReader$ParallelTermEnum.next(ParallelReader.java)
        at org.apache.lucene.search.DateFilter.bitsForField(DateFilter.java:)
        at org.apache.lucene.search.DateFilter.bits(DateFilter.java)
                
First of all, is it a good idea to get all the code from the trunk, instead of just pieces
to make ParallelReader work or is there a bug which I can just fix and move on with the existing
code ?
 
Question 2:
==========
My index 2 is expected to be dynamic, meaning will be deleted and reindexed often. Will this
not change the Doc ids? My understanding is, Parallel reader expects doc id (Lucene internal)
to be the same for it to query against these 2 indexes and return results appropriately. Is
my understanding wrong?
 
Appreciate any help in this respect.
 
Thank you
JS


		
---------------------------------
Yahoo! for Good
 Click here to donate to the Hurricane Katrina relief effort. 
Mime
  • Unnamed multipart/alternative (inline, 8-Bit, 0 bytes)
View raw message