Create new type `table`

[akariv]

This new JTS type will be identical to array, except it will have a mandatory attribute schema which will contain a valid JSON Table Schema for the items in the array.

For example:

"schema": {
  "fields": [
    { "name": "tender-id", "type": "string", "constraints": {"required": true} },
    { "name": "tender-documents",
      "type": "table",
      "schema": {
        "fields": [
          { "name": "document-date", "type": "date"},
          { "name": "document-title", "type": "string"},
          { "name": "document-url", "type": "string", "format": "uri"}
        ]
      }
    }
  ]
}

[rufuspollock]

@akariv i like the idea but sense this is post v1.0. Could you do a PR to add this as a pattern in the mean time?

[pwalsh]

+1 on this.

[pwalsh]

See my comment at #409 (comment)

Noting that this implementation maps well onto document-based data backends that support nesting (mongodb, Elasticsearch) and a version of this is already implemented in https://github.com/frictionlessdata/tableschema-elasticsearch-py