GeoAPI

 
100 of 309

Proposal for RecordType/Record implementation

Fixed

Description

Currently, RecordType is mapped to Class and Record is mapped to Object. This is fine on the surface, but does not permit the dynamic construction of arbitrary record types at runtime. The attached proposal fixes this by presenting an implementation of Figure 15 in ISO19103:2005(E). The interfaces "Schema", "RecordSchema", "Record" and "RecordType" all encapsulate dictionaries (Maps) which relate some form of LocalName to the indicated type. Figure 15 does not clearly indicate (to me) that the dictionary functionality should be publicly exposed, so I have included only the explicitly defined locate() method.

This Issue depends on [https://jira.codehaus.org/browse/GEO-71" title="Incorrect implementation of org.opengis.util.GenericName" class="issue-link" data-issue-key="GEO-71"><del>GEO-71</del>.

The Poseidon model attached to [https://jira.codehaus.org/browse/GEO-71" title="Incorrect implementation of org.opengis.util.GenericName" class="issue-link" data-issue-key="GEO-71"><del>GEO-71</del> contains this diagram as well.

Environment

None

Attachments

1

Details

Assignee

Reporter

Priority

Created March 30, 2015 at 12:20 AM
Updated May 26, 2015 at 2:45 PM
Resolved May 26, 2015 at 2:45 PM

Activity

codehaus 
April 10, 2015 at 4:51 PM

CodeHaus Comment From: desruisseaux - Time: Fri, 4 Sep 2009 10:13:46 -0500
---------------------
<p>The <tt>Record</tt> and related types were commited by Bryce a few years ago and have been quite stable. I did today a comparison with figure 15 in ISO 19103:2005. I added the <tt>Type</tt> interface as a safety, which was defined by ISO but not referenced in the original <tt>RecordType</tt> GeoAPI interface. This allow us to reference <tt>Type</tt> in the way ISO do, in a <tt>RecordType.getMemberTypes()</tt> method. With the addition of this change, the current interface seems to match raisonably well ISO 19103, so I'm closing this issue.</p>

codehaus 
April 10, 2015 at 4:51 PM

CodeHaus Comment From: desruisseaux - Time: Wed, 28 Nov 2007 14:06:34 -0600
---------------------
<p>This is documented in the OGC Change Request document and the interfaces should have been updated accordingly, but not yet tested in GeoTools implementation. This issue is kept open for that reason.</p>

codehaus 
April 10, 2015 at 4:51 PM

CodeHaus Comment From: jgarnett - Time: Fri, 13 Jan 2006 12:28:09 -0600
---------------------
<p>Hi Bryce, we got too alternatives - Deegree has Record extend Feature.</p>

<p>And we have both Feature and Record extend ComplexAttribute. The idea being that Expression works on Object (ie Attribute, Complex, Feature, Record) and can be used for to enable Filter to act as restricting "facets" for data validation.</p>

<p>Can I ask you review the feature model work and come back to this request ....</p>