Canvas Data 2 Nested Data Patterns

[patdura]

Will nested data always have the same pattern? 

Will nested data ever have deeply nested data (nested within nested data)? Other than Quiz tables which seem to have deeply nested data.

Can you confirm the schema of the nested data which is being sent as string?

Example 1: requirements_met column:

"[{"id":913185,"type":"must_view"},{"id":913186,"type":"min_score","min_score":80.0}]"

Will this always be an array of structs:

Array:

   struct:

      id:

      type:

       ....

Example 2: question_data column

"{"answers":[{"weight":100,"id":7482,"text":"Strongly agree","blank_id":"resources"},{"weight":0,"id":2552,"text":"Agree","blank_id":"resources"},{"weight":0,"id":2713,"text":"Neither agree nor disagree","blank_id":"resources"},{"weight":0,"id":7861,"text":"Disagree","blank_id":"resources"},{"weight":0,"id":8903,"text":"Strongly disagree","blank_id":"resources"},{"weight":0,"id":1367,"text":"Strongly agree","blank_id":"resources"},{"weight":0,"id":4647,"text":"Agree","blank_id":"resources"},{"weight":0,"id":4658,"text":"Neither agree nor disagree","blank_id":"resources"},{"weight":0,"id":5426,"text":"Disagree","blank_id":"resources"},{"weight":0,"id":1674,"text":"Strongly disagree","blank_id":"resources"}],"correct_comments":"","incorrect_comments":"","question_text":"<p>Please rate your agreement with the following statements.</p>\n<table style=\"height: 295px; width: 95%; border-collapse: collapse; border-style: solid;\" border=\"1\">\n<tbody>\n<tr style=\"height: 29px;\">\n<td style=\"width: 50%; height: 29px;\">Accessing the learner labs was easy</td>\n<td style=\"width: 50%; height: 29px;\">[accessing labs]</td>\n</tr>\n<tr style=\"height: 29px;\">\n<td style=\"width: 50%; height: 29px;\">The … assignment</td>\n<td style=\"width: 50%; height: 35px;\">[resources]</td>\n</tr>\n<tr style=\"height: 35px;\">\n<td style=\"width: 50%; height: 35px;\">The learner labs allowed me to practice skills I could not have otherwise</td>\n<td style=\"width: 50%; height: 29px;\">[practice skills]</td>\n</tr>\n<tr style=\"height: 29px;\">\n<td style=\"width: 50%; height: 29px;\">The … adding value to my learning experience</td>\n<td style=\"width: 50%; height: 29px;\">[added value]</td>\n</tr>\n</tbody>\n</table>","question_type":"multiple_dropdowns_question","question_name":"Question 1","migration_id":"mastercourse_15691_193_0bc7a7527db1afe986d24c1eef022029","question_bank_name":"Unfiled Questions","points_possible":0.0,"prepped_for_import":true}"

Will this answers(struct) always be an array of struct?

Struct:
  Array:
     Struct:
        weight:
        id:
        text:
        ....

[LeventeHunyadi]

> Will nested data always have the same pattern?

Depends.

When a column is declared as having a type of free text, DAP/CD 2 does not guarantee any particular inner structure. For example, the column requirements_met in the table context_module_progressions is declared as Annotated[str, ...], which means that it is free text (a string), and the structure may vary from row to row.

However, when the column is declared with a well-defined schema, e.g. the column question_data in the table quiz_questions, the internal structure is guaranteed, and forward and backward compatibility rules apply. Note that DAP follows the convention of omitting a property when its value is null, so you may not see every property in every object if they are marked as optional (nullable). For example, answers in quiz_questions__question_data has the type str | None, i.e. it is nullable.