Repository navigation
SPIKE-4: Document what ISO fields MDTranslator preserves and drops #6039
Description
What: The ISO-to-DCAT-US crosswalk embedded in MDTranslator determines which ISO 19115 fields survive the conversion. This crosswalk may not have been formally documented as a decision record for Data.gov. It is not clear which fields are intentionally dropped vs. accidentally missed, and agencies and downstream consumers (including GeoPlatform) have no reference for what to expect.
Using the Census concity record as one of many test cases, map the full MDTranslator output against the source ISO XML to produce a field-by-field accounting.
Investigate and document:
- Run a test case concity record through MDTranslator and capture the full output JSON
- For each ISO field in the source record, note whether it appears in the output, is absent, or is transformed/renamed
- Pay particular attention to:
- CRS (
MD_ReferenceSystem/referenceSystemInfo) - Spatial representation type (
MD_SpatialRepresentationTypeCode) - Scale denominator (
MD_Resolution/equivalentScale) - Topic category (
MD_TopicCategoryCode) - Data quality reports (
DQ_DataQuality/ report elements) - Lineage processing steps (
LI_Lineage/processStep) - Feature catalog link (
MD_FeatureCatalogueDescription) - Purpose (
gmd:purpose) - Status (
MD_ProgressCode)
- CRS (
- Categorize each dropped field as:
- (A) Intentionally out of scope -- no DCAT-US equivalent
- (B) Possible mapping exists in DCAT-US v3.0 -- not yet implemented
- (C) MDTranslator gap -- DCAT-US v1.1 field exists but MDTranslator does not output it
Done When:
- Field mapping table produced and attached to this ticket
- Each dropped field categorized A, B, or C
- Category B and C fields become candidates for IMPL-6 scope
Activity
rshewitt commentedon Jun 5, 2026
ISO records contain an enormous amount of information and can be assembled in many different combinations so the answer to this isn't simple. here's an example of an ISO-to-DCATUS1.1 crosswalk. The answer would be the total number of simple elements (those which contain values) in the doc minus those which are used in the output.
It's up to the people who create the crosswalks to consider the information available to them and make the right decision to map that to the output. Transformation of feature-rich schemas like the ISO19115 series is iterative where we continually pull more information.