Mailing-List: contact issues-help@drill.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@drill.apache.org
Date: Fri, 27 Mar 2015 17:25:54 +0000 (UTC)
From: "Daniel Barclay (Drill) (JIRA)" <jira@apache.org>
To: issues@drill.apache.org
Message-ID: <JIRA.12786261.1427476495000.51806.1427477154065@Atlassian.JIRA>
In-Reply-To: <JIRA.12786261.1427476495000@Atlassian.JIRA>
References: <JIRA.12786261.1427476495000@Atlassian.JIRA>
 <JIRA.12786261.1427476495402@arcas>
Subject: [jira] [Updated] (DRILL-2603) Review, resolve NULL
 checks/isNull()/null around vectors, accessors, JDBC SQLAccessor, etc.
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/DRILL-2603?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Daniel Barclay (Drill) updated DRILL-2603:
------------------------------------------
    Description: 
Review the checking for SQL NULL and mapping to Java null (for non-primitive types) from the vector and vector access layer up through the JDBC layer (e.g., SQLAccessors (which is just for JDBC)), resolve where checking should be done, and specify that contract information in the Javadoc documentation.

Considerations:
- With DRILL-2463, there are now some redundant checks for SQL NULL.
- Get methods returning non-primitive types must follow a different pattern, since they can't return Java null.
- Behavior that is very specific to JDBC usually should be implemented in the JDBC layer.  (For example, JDBC's NULL-reporting protocol of calling getInt(...) and then calling wasNull() shouldn't appear in lower layers (e.g., the planned DrillClient-level public interface) unless we intentionally choose that protocol for there.

  was:


Review NULL checks/null returns around vectors, accessors, JDBC SQLAccessor, etc.


> Review, resolve NULL checks/isNull()/null around vectors, accessors, JDBC SQLAccessor, etc.
> -------------------------------------------------------------------------------------------
>
>                 Key: DRILL-2603
>                 URL: https://issues.apache.org/jira/browse/DRILL-2603
>             Project: Apache Drill
>          Issue Type: Task
>            Reporter: Daniel Barclay (Drill)
>
> Review the checking for SQL NULL and mapping to Java null (for non-primitive types) from the vector and vector access layer up through the JDBC layer (e.g., SQLAccessors (which is just for JDBC)), resolve where checking should be done, and specify that contract information in the Javadoc documentation.
> Considerations:
> - With DRILL-2463, there are now some redundant checks for SQL NULL.
> - Get methods returning non-primitive types must follow a different pattern, since they can't return Java null.
> - Behavior that is very specific to JDBC usually should be implemented in the JDBC layer.  (For example, JDBC's NULL-reporting protocol of calling getInt(...) and then calling wasNull() shouldn't appear in lower layers (e.g., the planned DrillClient-level public interface) unless we intentionally choose that protocol for there.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)