Attribute definitions: a dictionary beside the values
Deliver field meanings, units and transformation rules beside the values. A Saputo reference and a fictional example of supported conversion.
Inriver’s written Saputo case describes product data from multiple regions brought into a consistent model. The provider reports standardizing product attributes across divisions, spanning regulatory, marketing and production needs, and delivering content for multiple commerce channels. These are Inriver’s published project statements. The page does not publish Saputo’s field dictionary, unit-conversion rules or multilingual value mappings. The following dictionary design is a PandaBuyNow editorial proposal for a separate marketplace-data delivery.
Define the field before normalizing its value
For a JSON or CSV export, PandaBuyNow proposes a companion dictionary with a stable field name, plain-language meaning, data type and the object it describes. Identify whether the value belongs to a listing, a selected variant or a package. Keep any permitted unit, value list and language convention with the definition.
Retain the source field label and original value in the record. A normalized value needs its transformation rule and rule version. A translated display label can share one field identifier while retaining its own language code. If the source does not establish the meaning, preserve that uncertainty in a data-quality note.
Test the dictionary with fictional data rows
Consider invented variant FRAME-D. Its source states width as “300 mm.” The recipient’s dictionary defines width_cm as the width of that selected variant, represented as a number in centimetres. An agreed conversion rule can produce 30 while preserving “300 mm,” the source reference and the rule version.
A second row states only “300,” with no unit. Leave its normalized width unresolved and include the missing-unit reason in the quality notes. Check that neither row is describing package width before combining them. These invented labels and values describe no Saputo record and no completed PandaBuyNow customer delivery.
Review the dictionary and file as one delivery
Ask the recipient to interpret a supported conversion, a missing unit and a translated field label using the delivered dictionary. Confirm that numeric types, permitted values and missing-value treatment survive the intended JSON or CSV format. Keep the dictionary version and data-quality notes with the file so another reader can reproduce the interpretation.
For a PandaBuyNow discussion, provide representative marketplace links, required fields and the recipient’s expected structure. Agree feasible extraction, transformation rules, source references and quality-note format through a small data excerpt. The output supports the customer’s data review and downstream use; unresolved source details remain documented for the customer to decide how to handle.