Defining Data Validation

Data validations can be defined using JSON. The ‘editChecks’ section of the code must be inserted after the ‘dictionary’ section in the workflow JSON.

The code has three parts - records, forms and rules
1. records - Whenever any of these specified records are modified; the constraint rules should be evaluated/enforced. Refers to the fields at the data entry level for which the validation is to be applied. The value for records can be cpr, specimen, visit or a combination of these.
2. forms - Optional field. If present, specifies the list of forms whose data is used to enforce the constraints.
3. rules - List of integrity rules that should be satisfied by the records and forms.
A ‘rule’ is made up of three attributes when, expr, and description.
1. when: The attribute 'when' is optional and specifies when the rule is applicable. For example, the rule is applicable only for primary specimens
2. expr: The attribute 'expr' specifies the check constraint that should be satisfied.
3. description: The attribute 'description' describes the constraint in user-friendly language. This is also used in error messages when the rule is broken or not satisfied.
Aliases for rules are - “cpr, visit, primarySpecimen, specimen”, which are self-explanator.
Aliases for accessing the form records are cprForms, visitForms, primarySpecimenForms, specimenForms. These are maps that can be indexed by the form name to access the relevant form data. For example, specimenForms['SpecimenFrozenEvent'] can be used to access the latest frozen event. Similarly, specimenForms['SpecimenFrozenEvent$Array'] can be used to access the list of all the frozen events for the specimen in question.

Example: Visit date should be same or later than the registration date

For more examples, please refer to the wiki page Data Validation Examples

{
  "name" : "editChecks",
  "data" : {
    "constraints" : [ 
      {
        "records" : [ "cpr", "visit" ],
        "rules" : [
          {
            "expr" : "#cpr.registrationDate != null && #visit.visitDate != null && !#cpr.registrationDate.after(#visit.visitDate)",
            "description" : "Visit date should be same or later than the registration date!"
          } 
        ]
      }
    ]
  }
}

Data Validation Fields

We can set up data validation on these 3 types of fields and their combinations:

Core fields
Custom fields
Custom form fields
Combination of the core fields, custom fields, custom form fields

Level	Field Name	Field Expression

Level	Field Name	Field Expression
Participant	PPID	`#cpr.ppid`
Participant	First Name	`#cpr.participant.firstName`
Participant	Middle Name
Participant	Last Name
Participant	Registration Date
Participant	External Subject ID
Participant	Registration Site
Participant	Birth Date
Participant	Social Security Number
Participant	eMPI
Participant	Gender
Participant	Vital Status
Participant	Death Date
Participant	Races
Participant	Ethnicity
Participant	MRN
Visit	Name
Visit	Name
Visit	Status
Visit	Missed By
Visit	Missed Reason
Visit	Visit Date
Visit	Visit Site
Visit	Clinical Diagnosis
Visit	Clinical Status
Visit	SPR
Visit	Visit Comments
Specimen	Label
Specimen	Barcode
Specimen	Lineage
Specimen	Collection Status
Specimen	Type
Specimen	Anatomic Site
Specimen	Laterality
Specimen	Pathology Status
Specimen	Initial Qty
Specimen	Available Qty
Specimen	Concentration
Specimen	Parent Specimen Label
Specimen	Biohazard
Specimen	Location
Specimen	Created On
Specimen	Freezer Thaw Cycle
Specimen	Increment Freeze Thaw Cycle
Specimen	Comments
Specimen	Collection Date
Specimen	Collector
Specimen	Received Date
Specimen	Receiver
Specimen	Collection Container
Specimen	Collection Procedure
Specimen	Received Quality
Specimen	Frozen Time
Specimen	Frozen Method
Specimen	Frozen Comments

Adding Multiple Data Validations

In case there are many data validations to be performed, here is a tip that could help in avoiding errors:

Check and test the data entry workflow after adding each validation. This will allow you to identify if any validation fails and will be useful in rectifying the failed one. If all the validation are added together, it will be difficult to identify the validation in case the above error ‘Cannot index into a null value’ occurs.