Difference beteween standard and JSON schema (fieldsMatch)

[dtemkin-volpe]

In the Table Schema documentation, fieldsMatch is described as "MUST be a string with the following possible values and the exact value by default".

However, in the JSON schema for the Table Schema, fieldsMatch is typed as an array as follows:

"fieldsMatch": {
            "type": "array",
            "item": {
                "type": "string",
                "enum": [
                    "exact",
                    "equal",
                    "subset",
                    "superset",
                    "partial"
                ],
                "default": "exact"
            }
        }

It seems to me like the standard describes fieldMatch as a string, but the JSON schema describes it as an array with a singular string item? I'm not really sure which one to follow (or if this is actually an issue, maybe I'm just reading everything wrong), so let me know. Thanks!

[peterdesmet]

Thanks for reporting, that seems to be an error in the JSON schema. fieldMatch needs to be a string, with an enum of possible values.

@roll I think this needs to be updated to:

 tableSchemaFieldsMatch: 
   type: string # <-- updated
   item: # <-- no need to define type here
     enum:
       - exact 
       - equal 
       - subset 
       - superset 
       - partial 
     default: exact 

That is similar to itemType:

datapackage/profiles/source/dictionary/schema.yaml

Lines 1094 to 1105 in cc20b49

	itemType:
	title: Item type.
	description: Specifies the list item type.
	type: string
	enum:
	- string
	- number
	- integer
	- boolean
	- datetime
	- date
	- time