cayenne-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrus Adamchik (JIRA)" <j...@apache.org>
Subject [jira] Updated: (CAY-1136) "Compiled" EntityResult ; Switch SelectQuery processing to use EntityResult
Date Fri, 21 Nov 2008 06:17:12 GMT

     [ https://issues.apache.org/cayenne/browse/CAY-1136?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Andrus Adamchik updated CAY-1136:
---------------------------------

    Summary: "Compiled" EntityResult ; Switch SelectQuery processing to use EntityResult 
(was: "Compiled" EntityResult ; Switch SelectQuery processign to use EntityResult)

> "Compiled" EntityResult ; Switch SelectQuery processing to use EntityResult
> ---------------------------------------------------------------------------
>
>                 Key: CAY-1136
>                 URL: https://issues.apache.org/cayenne/browse/CAY-1136
>             Project: Cayenne
>          Issue Type: Improvement
>          Components: Cayenne Core Library
>    Affects Versions: 3.0
>            Reporter: Andrus Adamchik
>            Assignee: Andrus Adamchik
>             Fix For: 3.0
>
>
> Related to the inheritance tasks (CAY-795)... We need a way to process all different
forms of DataRows that inheritance select queries require. Till now Cayenne worked with DataRows
that represented a single entity and used DbAttribute names as keys to the values. 
> There were a few exceptions that were handled by "normalization" process during object
creation. I want this to exception to become a rule, and abstract normalization process via
EntityResult, "compiled" in the ClassDescriptor. So for the old style rows , normalization
will be a noop (encapsulated in an EntityResult), and for the rows using special column naming
and extra meta-columns (such as in case of horizontal inheritance), there will be a conversion
process involved.
> Also the plan is to use EntityResult in SELECT ... clause column name generation, including
UNION queries under horizontal inheritance.
> [Finally going forward (will not be done under this Jira, or maybe not even in 3.0),
we may migrate to array-based positional DataRows instead of a current Map-based. Per CAY-999
this should provide huge memory savings, and preclude the need for "normalization".] 

-- 
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