Mailing-List: contact dev-help@atlas.incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@atlas.incubator.apache.org
Date: Tue, 2 May 2017 17:43:04 +0000 (UTC)
From: "Christopher Grote (JIRA)" <jira@apache.org>
To: dev@atlas.incubator.apache.org
Message-ID: <JIRA.13065457.1492680728000.105125.1493746984142@Atlassian.JIRA>
In-Reply-To: <JIRA.13065457.1492680728000@Atlassian.JIRA>
References: <JIRA.13065457.1492680728000@Atlassian.JIRA> <JIRA.13065457.1492680728654@jira-lw-us.apache.org>
Subject: [jira] [Commented] (ATLAS-1740) Amend types to be more intuitive
 when creating multiple cardinality attributes
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Tue, 02 May 2017 17:43:08 -0000


    [ https://issues.apache.org/jira/browse/ATLAS-1740?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15993374#comment-15993374 ] 

Christopher Grote commented on ATLAS-1740:
------------------------------------------

I agree with [~davidrad] -- this is confusing.

As another alternative to consider (though with much broader ramifications), could it be worthwhile to adopt something like JSON Schema (http://json-schema.org), or some similar definition language?  This might give broader consistency and recognition in terms of how the different characteristics are defined and understood.

In this example, we'd have:
{code:javascript}
{
  "type": "array",
  "items": {
    "$ref": "#/definitions/hive_column"
  }
}
{code}

> Amend types to be more intuitive when creating multiple cardinality attributes 
> -------------------------------------------------------------------------------
>
>                 Key: ATLAS-1740
>                 URL: https://issues.apache.org/jira/browse/ATLAS-1740
>             Project: Atlas
>          Issue Type: Bug
>            Reporter: David Radley
>
> When specifying a multiple carnality attribute - for example hive-columns , Atlas has 
>                     "typeName": "array<hive_column>",
>                     "cardinality": "SINGLE",
> This is confusing - I read this as a single array. 
> I would expect MULTIPLE cardinality to be used. A more intuitive way would be to use:
>                     "typeName": "hive_column",
>                     "cardinality": "MULTIPLE",
> How we migrate the existing API should be considered here. 
> I suggest clear design around when to use arrays and cardinality and what the difference - if any - is. 
>  
>  


--
This message was sent by Atlassian JIRA
(v6.3.15#6346)