Open philipashlock opened 8 years ago
To date this had been suggested (and documented in the crosswalk) using the following fields in ISO 19115 as equivalents to map from isPartOf
. It looks like others (e.g. here or here) have used parentIdentifier
to create parent/child relationships just like we use isPartOf
in the Project Open Data Metadata Schema.
//gmd:identificationInfo/gmd:MD_DataIdentification/gmd:aggregationInfo/gmd:MD_AggregateInformation/gmd:aggregateDataSetIdentifier/gmd:MD_Identifier/gmd:code
(gco:CharacterString or gmx:Anchor)
+
//gmd:identificationInfo/gmd:MD_DataIdentification/gmd:aggregationInfo/gmd:MD_AggregateInformation/gmd:associationType/gmd:DS_AssociationTypeCode == largerWorkCitation
//gmd:identificationInfo/gmd:MD_DataIdentification/gmd:aggregationInfo/gmd:MD_AggregateInformation/gmd:aggregateDataSetName/gmd:CI_Citation/gmd:title
+
//gmd:identificationInfo/gmd:MD_DataIdentification/gmd:aggregationInfo/gmd:MD_AggregateInformation/gmd:associationType/gmd:DS_AssociationTypeCode == largerWorkCitation
//gmd:parentIdentifier
(gco:CharacterString or gmx:Anchor)For reference: http://schemas.opengis.net/
Summary of resolution as documented in 'Geospatial Metadata Recommendations Supporting Discovery and Use in Data.gov and the Geospatial Platform' vers. 20170302 (pending publication to https://cms.geoplatform.gov/resources)
Below are suggestions from Geo-spatial team on metadata vs data resource identifiers for consideration
Metadata vs Data (resource) Identifiers
Metadata Record Identifiers MD_Metadata elements are specific to the metadata record ISO 19115 (these are the elements that Data.gov is currently using - only fileIdentifier) MD_Metadata.fileIdentifier - Metadata record of resource being documented MD_Metadata.parentIdentifier - Metadata record from which the current metadata record is a ‘subset’ ISO 19115-1 (note names have been changed on both to make relationship more clear) MD_Metadata.metadataIdentifier MD_Metadata.parentMetadata
A metadata record can be the child of another metadata record without being a child of the resource (dataset, service, etc.) that the metadata record represents.
For example, a metadata record may be the child (subset) of an organizational metadata record (parent) that contains contacts, liability and constraint language, distribution methods, QA/QC methods, etc.). It is good data management practice to create subset metadata records and separate fixed (parent) and dataset-specific (child) metadata and compile upon publication. In this way, fixed information can be updated in one place (parent metadata record) vs. each associated (child) metadata record.
Data (resource) Identifiers MD_Identification elements are specific to the resource being documented (data, service, etc.), e.g. Resource Identifier MD_Identification>Citation.identifier Parent (or other association) Resource Identifier 19115: MD_Idenfication.aggregationInfo>MD_AggregateInformation. aggregateDataSetName>Citation.identifier 19115-1: MD_Idenfication.associatedResource>MD_AssociatedResource.name
Citation.identifier And MD_AssociatedResource.type (LargerWorkCitation, partOfSeamlessDatabase, collectiveTitle, series)
NOTE that type = is ComposedOf can be used to indicate child resources
If the Data.gov objective is to identify resources that have parent/.child, and other relations, e.g. a data collection and a member of that collection, then the MD_Identification>Citation.identifier would be the more appropriate identifier to document that relation.
May be that the MD_Metadata.fileIdentifier may be best to uniquely identify the metadata record and the resource identifiers used to document parent/child relations between resources.
Currently collections can be defined within metadata using the Project Open Data Metadata schema, but this is not currently supported with geospatial metadata. Currently the only way geospatial metadata can be organized into collections is if the children in the collection have their own separate WAF harvest source with that harvest source manually registered as a collection.