db-ddlutils-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tom Palmer (JIRA)" <j...@apache.org>
Subject [jira] Updated: (DDLUTILS-258) Degradation of performance of Table when using DdlUtils with a large dataset
Date Sun, 25 Apr 2010 18:03:50 GMT

     [ https://issues.apache.org/jira/browse/DDLUTILS-258?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Tom Palmer updated DDLUTILS-258:
--------------------------------

    Attachment: ListOrderedSet.patch

The attached patch changes the _columns, _foreignKeys and _indicies fields to use the Commons
Collections ListOrderedSet. This allows the order to be preserved as well as ensuring uniqueness
of the elements. With this change I've been able to remove the 4x HashSet creations in equals
and hashCode. sortForeignKeys must now create a temporary ArrayList to perform the sorting
but this is called far less often so I don't believe this to be a problem. 

> Degradation of performance of Table when using DdlUtils with a large dataset
> ----------------------------------------------------------------------------
>
>                 Key: DDLUTILS-258
>                 URL: https://issues.apache.org/jira/browse/DDLUTILS-258
>             Project: DdlUtils
>          Issue Type: Improvement
>          Components: Core (No specific database)
>    Affects Versions: 1.1
>            Reporter: Tom Palmer
>            Assignee: Thomas Dudziak
>            Priority: Minor
>         Attachments: ListOrderedSet.patch
>
>
> When using DdlUtils to load large amounts of data I've observed a drop-off in performance
originating in the Table class. It uses an ArrayList to store its columns, foreign keys and
indicies but treats the contents as a Set within the equals, hashCode and sortForeignKeys.
On one particular project I've observed up to 2 minutes of the overall load time being consumed
in these methods.

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


Mime
View raw message