Problem to Solve
The current extraction engine only supports scalar values (STRING, DECIMAL) as targetType in searchFields. There is no way to extract fields whose value is an array of objects, even when the number of elements is unknown at extraction time.
Proposed Solution
The extraction configuration should support a new targetType such as ARRAY or OBJECT_ARRAY, allowing the extractor to capture all elements of a typed array and store them as a JSON array in search_data.
Example proposed config:
{
"fieldName": "materialsProvenance",
"targetType": "ARRAY"
}
With a corresponding strategy entry (e.g. in unknownOntology):
"materialsProvenance": {
"typeHints": ["Material"],
"collectAll": true
}
The resulting search_data entry would be:
"materialsProvenance": [
{ "materialName": "...", "originCountry": "...", "concentration": 100 },
{ "materialName": "...", "originCountry": "...", "concentration": 100 }
]
Priority
High (blocking for my use case)
Use Case
When extracting data from a Battery Digital Product Passport, some fields could be an array of objects:
example:
"dpp:materialsProvenance": [
{
"@type": "dpp:Material",
"dpp:materialName": "...",
"dpp:originCountry": "...",
"dpp:certifiedContent": true,
"dpp:materialType": "...",
"dpp:concentration": 100
},
{
"@type": "dpp:Material",
"dpp:materialName": "...",
"dpp:originCountry": "...",
"dpp:certifiedContent": false,
"dpp:materialType": "...",
"dpp:concentration": 100
}
]
The number of elements in the array is unknown at configuration time and can vary per DPP. There is currently no targetType or extraction strategy that can handle this structure.
Workaround
Currently the only workaround is to serialize the array as a concatenated STRING at the provider level, which loses structure and makes filtering/rendering on the client side much harder.
Technical Notes (optional)
- Extractor version: `1.0.0-pgsql-oidc`
- DPP format: JSON-LD with `@context` referencing UNTP/DPPO/BattINFO vocabularies
- Strategy affected: `unknownOntology` (and likely `noOntology`)
Problem to Solve
The current extraction engine only supports scalar values (
STRING,DECIMAL) astargetTypeinsearchFields. There is no way to extract fields whose value is an array of objects, even when the number of elements is unknown at extraction time.Proposed Solution
The extraction configuration should support a new
targetTypesuch asARRAYorOBJECT_ARRAY, allowing the extractor to capture all elements of a typed array and store them as a JSON array insearch_data.Example proposed config:
{ "fieldName": "materialsProvenance", "targetType": "ARRAY" }With a corresponding strategy entry (e.g. in
unknownOntology):The resulting
search_dataentry would be:Priority
High (blocking for my use case)
Use Case
When extracting data from a Battery Digital Product Passport, some fields could be an array of objects:
example:
The number of elements in the array is unknown at configuration time and can vary per DPP. There is currently no
targetTypeor extraction strategy that can handle this structure.Workaround
Currently the only workaround is to serialize the array as a concatenated
STRINGat the provider level, which loses structure and makes filtering/rendering on the client side much harder.Technical Notes (optional)