You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucenenet.apache.org by "Nitin Shiralkar (JIRA)" <ji...@apache.org> on 2008/02/06 08:21:08 UTC
[jira] Created: (LUCENENET-109) Severe search performance drop in
case of sorting on ID fields
Severe search performance drop in case of sorting on ID fields
---------------------------------------------------------------
Key: LUCENENET-109
URL: https://issues.apache.org/jira/browse/LUCENENET-109
Project: Lucene.Net
Issue Type: Bug
Reporter: Nitin Shiralkar
Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
Lucene Index:
1. Document count: ~1000,000
2. Index size: > 8 GB
Search Query:
document_subtype:1.1201
Lucene Search API:
topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
Re: [jira] Created: (LUCENENET-109) Severe search performance drop in case of sorting on ID fields
Posted by Jokin Cuadrado <jo...@gmail.com>.
could you provide more information?
- which version of lucene.net you use
- how many clauses have the query
- how many time took the query without search
- if you have read and applied all the points in the lucene
performance faq (it's the one of the java version, but applies
completely to the .net version)
- do you optimize the index before doing the query
- with how much frequency you reopen the indexreader.
without that is very difficult to make a guess of what can be happen.
--
Jokin
On Feb 6, 2008 8:21 AM, Nitin Shiralkar (JIRA) <ji...@apache.org> wrote:
> Severe search performance drop in case of sorting on ID fields
> ---------------------------------------------------------------
>
> Key: LUCENENET-109
> URL: https://issues.apache.org/jira/browse/LUCENENET-109
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Nitin Shiralkar
>
>
> Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
>
> Lucene Index:
> 1. Document count: ~1000,000
> 2. Index size: > 8 GB
>
> Search Query:
> document_subtype:1.1201
>
> Lucene Search API:
> topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
>
>
> --
> This message is automatically generated by JIRA.
> -
> You can reply to this email to add a comment to the issue online.
>
>
[jira] Commented: (LUCENENET-109) Severe search performance drop in
case of sorting on ID fields
Posted by "Digy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENENET-109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12566178#action_12566178 ]
Digy commented on LUCENENET-109:
--------------------------------
To be more specific (as Jokin suggests)
http://wiki.apache.org/lucene-java/ImproveSearchingSpeed
DIGY
> Severe search performance drop in case of sorting on ID fields
> ---------------------------------------------------------------
>
> Key: LUCENENET-109
> URL: https://issues.apache.org/jira/browse/LUCENENET-109
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Nitin Shiralkar
>
> Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
> Lucene Index:
> 1. Document count: ~1000,000
> 2. Index size: > 8 GB
> Search Query:
> document_subtype:1.1201
> Lucene Search API:
> topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (LUCENENET-109) Severe search performance drop in
case of sorting on ID fields
Posted by "Digy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENENET-109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12566158#action_12566158 ]
Digy commented on LUCENENET-109:
--------------------------------
This happens mostly when the indexreader(indexsearcher) is reopened with every search.
DIGY
> Severe search performance drop in case of sorting on ID fields
> ---------------------------------------------------------------
>
> Key: LUCENENET-109
> URL: https://issues.apache.org/jira/browse/LUCENENET-109
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Nitin Shiralkar
>
> Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
> Lucene Index:
> 1. Document count: ~1000,000
> 2. Index size: > 8 GB
> Search Query:
> document_subtype:1.1201
> Lucene Search API:
> topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (LUCENENET-109) Severe search performance drop in
case of sorting on ID fields
Posted by "Nitin Shiralkar (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENENET-109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12566931#action_12566931 ]
Nitin Shiralkar commented on LUCENENET-109:
-------------------------------------------
Ok. We have some design issues in our system not allowing us to use the same instance. But then why is that not a problem while sorting on other fields? Is there any other workaround available?
> Severe search performance drop in case of sorting on ID fields
> ---------------------------------------------------------------
>
> Key: LUCENENET-109
> URL: https://issues.apache.org/jira/browse/LUCENENET-109
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Nitin Shiralkar
>
> Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
> Lucene Index:
> 1. Document count: ~1000,000
> 2. Index size: > 8 GB
> Search Query:
> document_subtype:1.1201
> Lucene Search API:
> topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Closed: (LUCENENET-109) Severe search performance drop in
case of sorting on ID fields
Posted by "Digy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENENET-109?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Digy closed LUCENENET-109.
--------------------------
Resolution: Won't Fix
Assignee: Digy
see
http://wiki.apache.org/lucene-java/ImproveSearchingSpeed
> Severe search performance drop in case of sorting on ID fields
> ---------------------------------------------------------------
>
> Key: LUCENENET-109
> URL: https://issues.apache.org/jira/browse/LUCENENET-109
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Nitin Shiralkar
> Assignee: Digy
>
> Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
> Lucene Index:
> 1. Document count: ~1000,000
> 2. Index size: > 8 GB
> Search Query:
> document_subtype:1.1201
> Lucene Search API:
> topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
[jira] Commented: (LUCENENET-109) Severe search performance drop in
case of sorting on ID fields
Posted by "Digy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENENET-109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12566954#action_12566954 ]
Digy commented on LUCENENET-109:
--------------------------------
You can sort your results by yourself (after a lucene search without sorting).
This issue is related with the warming up of the indexreader and there is already an open issue LUCENENET-107.
So,it is better to close this issue and continue the discussion on mailing lists.
DIGY
> Severe search performance drop in case of sorting on ID fields
> ---------------------------------------------------------------
>
> Key: LUCENENET-109
> URL: https://issues.apache.org/jira/browse/LUCENENET-109
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Nitin Shiralkar
>
> Severe search performance drop in case of sorting on ID fields. I am using a lucene index where if I query for documents (500 result set) with sorting on document date field, then I get a resposne in ~2000 ms. However for the same query with sorting on Document Id field, then I get a response in ~4500 ms. This drop in search performance is even more severe on larger index.
> Lucene Index:
> 1. Document count: ~1000,000
> 2. Index size: > 8 GB
> Search Query:
> document_subtype:1.1201
> Lucene Search API:
> topDocs = indexSearcher.Search (userQuery, null, m_nMaximumRecordSetSize, sortField );
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.