You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Jonathan Ellis (JIRA)" <ji...@apache.org> on 2011/04/29 17:26:03 UTC

[jira] [Created] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

optimize batches containing multiple updates to the same rows
-------------------------------------------------------------

                 Key: CASSANDRA-2583
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
             Project: Cassandra
          Issue Type: Improvement
          Components: API
    Affects Versions: 0.8 beta 1
            Reporter: Jonathan Ellis
            Assignee: Pavel Yaskevich
            Priority: Minor
             Fix For: 0.8.1


Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Commented] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13030630#comment-13030630 ] 

Jonathan Ellis commented on CASSANDRA-2583:
-------------------------------------------

needs rebase

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Commented] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13030915#comment-13030915 ] 

Jonathan Ellis commented on CASSANDRA-2583:
-------------------------------------------

are there already tests that update a single row multiple times in the same batch?

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583-v2.patch, CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Commented] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Pavel Yaskevich (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13030636#comment-13030636 ] 

Pavel Yaskevich commented on CASSANDRA-2583:
--------------------------------------------

835669aee8dc0ba14a37594e1ff07d056c74a3ea is git-svn-id: https://svn.apache.org/repos/asf/cassandra/trunk@1100900 13f79535-47bb-0310-9956-ffa450edef68, but I see that it depends on CASSANDRA-2476 so please apply this one after CASSANDRA-2476 if possible.

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583-v2.patch, CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Commented] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Pavel Yaskevich (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13030918#comment-13030918 ] 

Pavel Yaskevich commented on CASSANDRA-2583:
--------------------------------------------

Yes, there are tests where in BATCH we at first INSERT a row and then update it or even delete it.

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583-v2.patch, CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Updated] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Pavel Yaskevich (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Pavel Yaskevich updated CASSANDRA-2583:
---------------------------------------

    Attachment: CASSANDRA-2583.patch

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Commented] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13030634#comment-13030634 ] 

Jonathan Ellis commented on CASSANDRA-2583:
-------------------------------------------

hmm, still seeing failures

{noformat}
form:svn-trunk jonathan$ patch -p1 < CASSANDRA-2583-v2.patch 
patching file src/java/org/apache/cassandra/cql/AbstractModification.java
Hunk #2 succeeded at 86 (offset -7 lines).
Hunk #3 succeeded at 100 (offset -7 lines).
patching file src/java/org/apache/cassandra/cql/BatchStatement.java
Hunk #2 succeeded at 78 (offset -9 lines).
Hunk #3 succeeded at 121 (offset -9 lines).
patching file src/java/org/apache/cassandra/cql/DeleteStatement.java
patching file src/java/org/apache/cassandra/cql/UpdateStatement.java
Hunk #2 FAILED at 156.
1 out of 2 hunks FAILED -- saving rejects to file src/java/org/apache/cassandra/cql/UpdateStatement.java.rej
{noformat}

maybe git mirror doesn't have r1100900 yet?

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583-v2.patch, CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Updated] (CASSANDRA-2583) optimize batches containing multiple updates to the same rows

Posted by "Pavel Yaskevich (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-2583?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Pavel Yaskevich updated CASSANDRA-2583:
---------------------------------------

    Attachment: CASSANDRA-2583-v2.patch

rebased with the latest trunk (835669aee8dc0ba14a37594e1ff07d056c74a3ea)

> optimize batches containing multiple updates to the same rows
> -------------------------------------------------------------
>
>                 Key: CASSANDRA-2583
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-2583
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.8 beta 1
>            Reporter: Jonathan Ellis
>            Assignee: Pavel Yaskevich
>            Priority: Minor
>             Fix For: 0.8.1
>
>         Attachments: CASSANDRA-2583-v2.patch, CASSANDRA-2583.patch
>
>
> Currently we turn batches into one RowMutation per statement. This is suboptimal if we have multiple statements operating on the same row. This could happen either with generated code, or deliberately to allow different options to different columns in the row, e.g. for TTLs (CASSANDRA-2476).

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira