[jira] Created: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery

classic Classic list List threaded Threaded
69 messages Options
1234
Reply | Threaded
Open this post in threaded view
|

[jira] Created: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery

JIRA jira@apache.org
Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery
----------------------------------------------------------------------------------------------

                 Key: LUCENE-1771
                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
             Project: Lucene - Java
          Issue Type: Bug
          Components: Search
            Reporter: Mark Miller
            Assignee: Mark Miller
             Fix For: 2.9




--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery

JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12737288#action_12737288 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

reminder to add warning for custom queries - you shouldn't use top level reader to access fieldcache values for explain

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery
> ----------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Updated: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

     [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Mark Miller updated LUCENE-1771:
--------------------------------

    Summary: Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.  (was: Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery)

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Updated: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

     [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Mark Miller updated LUCENE-1771:
--------------------------------

    Attachment: LUCENE-1771.patch

Current state of fix (being iterated on in the fieldcache introspection issue - will be fully spun over to this issue after its shaken out)

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739106#action_12739106 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

Unless there is an objection, and after another review, I will commit this soon.

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739124#action_12739124 ]

Michael McCandless commented on LUCENE-1771:
--------------------------------------------

I think we should change the approach here, based on Yonik's last suggestion on LUCENE-1749, so that QueryWeight.explain receives *both* the top-level Searcher and the sub-reader, with the docID pre-resolved to that sub-reader.  This way individual queries are not responsible for resolving to sub-readers.

Actually, since QueryWeight is new in 2.9, we can simply make this change w/o deprecation.

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739126#action_12739126 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

okay, sounds good

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739130#action_12739130 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

searcher doesn't have access to numDocs (termqueryweight) at the moment ...

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Issue Comment Edited: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739130#action_12739130 ]

Mark Miller edited comment on LUCENE-1771 at 8/4/09 1:16 PM:
-------------------------------------------------------------

searcher doesn't have access to numDocs (termqueryweight) at the moment ...

*edit*

N/M - the IndexSearcher will have getIndexReader

      was (Author: [hidden email]):
    searcher doesn't have access to numDocs (termqueryweight) at the moment ...
 

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739132#action_12739132 ]

Yonik Seeley commented on LUCENE-1771:
--------------------------------------

bq. searcher doesn't have access to numDocs (termqueryweight) at the moment ...

Good... cause it's a bug and it shouldn't be using it ;-)  When was that added?

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739135#action_12739135 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

bq. Actually, since QueryWeight is new in 2.9, we can simply make this change w/o deprecation.

We do have to be back compat with Weight though - ugg - so QueryWeight would somehow need to expose a deprecated working explain(IndexReader, int) ? ...
but if we could do that ...

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739137#action_12739137 ]

Yonik Seeley commented on LUCENE-1771:
--------------------------------------

bq. Good... cause it's a bug and it shouldn't be using it  When was that added?

Tried out IntelliJ's "show history for selection" for the first time - cool stuff!
Anyway, the bug was introduced in LUCENE-1066, Nov-2007

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739138#action_12739138 ]

Yonik Seeley commented on LUCENE-1771:
--------------------------------------

bq. We do have to be back compat with Weight though - ugg - so QueryWeight would somehow need to expose a deprecated working explain(IndexReader, int) ?

QueryWeight doesn't need a explain(IndexReader, int)... but perhaps implementations of explain(IndexReader, Searcher, int) do need to handle a null searcher.  And no, the *output* of explain won't be 100% compatible (or 100% accurate) but we're too far down the road of per-segment searching, and too close to a release to fix that now IMO.

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739145#action_12739145 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

CustomScoreQuery has no access to the top level Searcher without docking it on the QueryWeight -

    /*(non-Javadoc) @see org.apache.lucene.search.Scorer#explain(int) */
    public Explanation explain(int doc) throws IOException {
      Explanation subQueryExpl = weight.subQueryWeight.explain(null, reader,doc); // nocommit: needs resolution

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739148#action_12739148 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

bq. QueryWeight doesn't need a explain(IndexReader, int)...


Why not? QueryWeight implements Weight for back compat ...

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739154#action_12739154 ]

Yonik Seeley commented on LUCENE-1771:
--------------------------------------

bq.  Why not? QueryWeight implements Weight for back compat...

Couldn't QueryWeight just have the following default implementation?
 public Explanation explain(IndexReader reader, int doc) throws IOException {
   explain(reader,doc,null);
 }

And implementations would just need to be able to handle a null searcher (for now).

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739160#action_12739160 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

Yes - but then I think we have to use the top reader for that (impls already expect it) - so you could have double caching problems where Weight is still used. I think we will have to live with that though.

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739166#action_12739166 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

bq. This way individual queries are not responsible for resolving to sub-readers.

... so who is? The user I guess? The only place explain(IndexReader reader, int doc) appears to be called in Lucene code is with the customscore query. So with the current patch, you can call these methods and pass the top level reader like normal, but with this change, you have to get the subreader yourself.

Also, the one internal place that  explain(IndexReader reader, int doc) is called (CustomScoreQuery) would have to pass null for the IndexSearcher.

Something doesn't seem right with this stuff ...

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739173#action_12739173 ]

Yonik Seeley commented on LUCENE-1771:
--------------------------------------

bq. .. so who is? (responsible for resolving to sub-readers)

I assumed IndexSearcher.explain().... which brings up another point... isn't a back-compat break too (since Weight is not a QueryWeight)?

IndexSearcher.Explanation explain(QueryWeight weight, int doc) throws IOException;



> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1771) Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12739182#action_12739182 ]

Mark Miller commented on LUCENE-1771:
-------------------------------------

bq. which brings up another point... isn't a back-compat break too (since Weight is not a QueryWeight)? IndexSearcher.Explanation explain(QueryWeight weight, int doc) throws IOException;

nice catch.

Hmmm - so scorer.explain(doc) would need to be deprecated cause it would need to take a Searcher (too avoid putting it on QueryWeight) and the sub-reader - and again this would need to be solved (like the current patch) a Query at a time.

> Using explain may double ram reqs for fieldcaches when using ValueSourceQuery/CustomScoreQuery or for ConstantScoreQuerys that use a caching Filter.
> ----------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: LUCENE-1771
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1771
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Search
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>             Fix For: 2.9
>
>         Attachments: LUCENE-1771.patch
>
>


--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

1234