Pending Approval

Pending Approval Mapping CF

CREATE TABLE IF NOT EXISTS pending_approval_mapping (
catchment_id varchar,
last_updated timeuuid,
health_id varchar,
PRIMARY KEY (catchment_id, last_updated)
)WITH CLUSTERING ORDER BY (last_updated ASC);

CREATE INDEX idx_pending_approval_mapping_health_id on pending_approval_mapping(health_id);

Search All Pending Approvals by Catchment

• health_ids and last_updated timeuuids are selected from pending_approval_mapping CF for a given catchment and gte timeuuid. The results are ordered by last_updates asc.

• For each health_id, patient CF is queried for given_name and surname. 

• Sample JSON response: 

{
   "hid": "5958385569250672641",
   "given_name": "Sony",
   "sur_name": "Bazzardo",
   "last_updated": "b385af50-a225-11e4-97c2-5fcb9978cb86"
}

Search by HID

• Pending approvals for a particular patient is stored in the patient CF. When pending approval for a particular patient (by HID) is queried, the result is served from patient CF.

• The field is 'pending_approvals.' The pending approvals are stored as a JSON string. Class PendingApproval is used to serialize/deserialize the JSON. Updates are grouped by fields.

• The pending approvals are ordered by field name (lexicographically), then timeuuid in desc order. For each update the timeuuid is captured. See 'field_details' in the JSON below.

• TODO: Spike user defined types in Cassandra

• Sample JSON response:

{
   "field_name": "gender",
   "current_value": "M",
   "field_details": {
      "9b450750-b28e-11e4-a7a6-5fcb9978cb86": {
         "value": "O",
         "facility_id": "10000059",
         "created_at": "2015-02-12T08:10:28.805Z"
   },
      "98bc87b0-b28e-11e4-a7a6-5fcb9978cb86": {
         "value": "F",
         "facility_id": "10000059",
         "created_at": "2015-02-12T08:10:24.555Z"
   }

}

Update Fields Marked for Approval

• When a particular patient field that is marked for approval is updated, patient CF and pending_approval_mapping CF are updated.

• patient CF - new field details added to the pending_approvals JSON.

• pending_approval_mapping CF - 

  • Search by health_id (secondary index) and delete the records. 

  • Create new records for each possible catchment (see explanation below). last_updated is set as the latest timeuuid of the pending_approvals in patient CF.

  • Possible catchments - If catchment is 10203040, the possible catchments are 1020, 102030, 10203040. Catchments should have always division and district.

• All queries run in a batch.

Approve Request

• The field(s) is updated in patient CF.

• When a field that is pending for approval is approved, all the updates for that field are removed from the pending_approvals in patient CF. 

• If there is no other pending_approvals in the patient CF, all pending_approval_mapping CF records for that health_id is deleted.

• Else, pending_approval_mapping CF  is updated as mentioned above.

• All queries run in a batch.

Reject Request

• The field(s) is NOT updated in patient CF.

• When a field that is pending for approval is rejected, only that particular updates for that field is removed from the pending_approvals in patient CF. 

• If there is no other pending_approvals in the patient CF, all pending_approval_mapping CF records for that health_id is deleted.

• Else, pending_approval_mapping CF  is updated as mentioned above.

• All queries run in a batch.