When we create the avro schemas for each extension to create the tables in iceberg we normalize the term names to handle duplicate names in extensions or to remove some characters like underscores. When we download those verbatim extensions we use these normalized names but I think we should use the original names defined in DWC.
Examples of the changes:
dna_sequence -> dnasequence
type from http://purl.org/dc/elements/1.1/ -> dc_type
type from http://purl.org/dc/terms/ -> dcterms_type
But the headers of the downloaded files should be:
dna_sequence
type
type
For the ones that are duplicated, the columns in the csv would have the same name but then you can see which term they refer to by looking at the meta.xml.
This normalization is also being the source of other errors such as #526 or gbif/pipelines#1408.
When we create the avro schemas for each extension to create the tables in iceberg we normalize the term names to handle duplicate names in extensions or to remove some characters like underscores. When we download those verbatim extensions we use these normalized names but I think we should use the original names defined in DWC.
Examples of the changes:
dna_sequence -> dnasequence
type from http://purl.org/dc/elements/1.1/ -> dc_type
type from http://purl.org/dc/terms/ -> dcterms_type
But the headers of the downloaded files should be:
dna_sequence
type
type
For the ones that are duplicated, the columns in the csv would have the same name but then you can see which term they refer to by looking at the meta.xml.
This normalization is also being the source of other errors such as #526 or gbif/pipelines#1408.