Migrate from record node UUID to internal ID

src/db/dbMigrator/migration/survey/migrations/20260217085512-alter-table-node-add-iid.js

@@ -0,0 +1,51 @@
+'use strict'
+
+var dbm
+var type
+var seed
+var fs = require('fs')
+var path = require('path')
+var Promise
+
+/**
+ * We receive the dbmigrate dependency from dbmigrate initially.
+ * This enables us to not have to rely on NODE_PATH.
+ */
+exports.setup = function (options, seedLink) {
+  dbm = options.dbmigrate
+  type = dbm.dataType
+  seed = seedLink
+  Promise = options.Promise
+}
+
+exports.up = function (db) {
+  var filePath = path.join(__dirname, 'sqls', '20260217085512-alter-table-node-add-iid-up.sql')
+  return new Promise(function (resolve, reject) {
+    fs.readFile(filePath, { encoding: 'utf-8' }, function (err, data) {
+      if (err) return reject(err)
+      console.log('received data: ' + data)
+
+      resolve(data)
+    })
+  }).then(function (data) {
+    return db.runSql(data)
+  })
+}
+
+exports.down = function (db) {
+  var filePath = path.join(__dirname, 'sqls', '20260217085512-alter-table-node-add-iid-down.sql')
+  return new Promise(function (resolve, reject) {
+    fs.readFile(filePath, { encoding: 'utf-8' }, function (err, data) {
+      if (err) return reject(err)
+      console.log('received data: ' + data)
+
+      resolve(data)
+    })
+  }).then(function (data) {
+    return db.runSql(data)
+  })
+}
+
+exports._meta = {
+  version: 1,
+}

src/db/dbMigrator/migration/survey/migrations/sqls/20260217085512-alter-table-node-add-iid-down.sql

@@ -0,0 +1 @@
+/* Replace with your SQL commands */

src/db/dbMigrator/migration/survey/migrations/sqls/20260217085512-alter-table-node-add-iid-up.sql

@@ -0,0 +1,118 @@
+ALTER TABLE node
+	ADD COLUMN i_id INTEGER;
+
+ALTER TABLE node
+	ADD COLUMN p_i_id INTEGER;
+
+CREATE INDEX IF NOT EXISTS node_uuid_idx 
+    ON node (uuid);
+
+CREATE INDEX IF NOT EXISTS node_id_idx 
+    ON node (id);
+
+CREATE INDEX IF NOT EXISTS node_record_uuid_id_idx 
+    ON node (record_uuid, id);
+
+-- Update node i_id START
+
+CREATE UNLOGGED TABLE temp_node_ranking AS
+SELECT 
+    id, 
+    ROW_NUMBER() OVER (PARTITION BY record_uuid ORDER BY id) AS new_i_id
+FROM node
+ORDER BY id;
+
+CREATE INDEX idx_temp_node_id ON temp_node_ranking (id);
+ANALYZE temp_node_ranking;
+
+SET LOCAL session_replication_role = 'replica';
+
+SET LOCAL work_mem = '512MB';
+
+UPDATE node
+SET i_id = t.new_i_id
+FROM temp_node_ranking t
+WHERE node.id = t.id;
+
+DROP TABLE temp_node_ranking;
+
+-- Update node i_id END
+
+ALTER TABLE node
+	ALTER COLUMN i_id SET NOT NULL;
+
+-- Populate the p_i_id column by joining the node table with itself based on the parent-child relationship
+UPDATE node AS child
+SET p_i_id = parent.i_id
+FROM node AS parent
+WHERE child.parent_uuid IS NOT NULL 
+    AND child.parent_uuid = parent.uuid;
+
+-- Replace meta.h UUID array with i_id array, preserving order
+UPDATE node AS n
+SET meta = jsonb_set(
+	meta,
+	'{h}',
+	COALESCE(
+		(
+			SELECT jsonb_agg(p.i_id ORDER BY elems.ordinality)
+			FROM jsonb_array_elements_text(n.meta->'h') WITH ORDINALITY AS elems(uuid_text, ordinality)
+			JOIN node AS p 
+				ON p.record_uuid = n.record_uuid 
+				AND p.uuid = elems.uuid_text::uuid
+		),
+		'[]'::jsonb
+	)
+)
+WHERE n.meta ? 'h';
+
+-- Update record validation fields to replace UUID references with i_id references
+
+UPDATE record r
+SET validation = jsonb_set(
+    r.validation, 
+    '{fields}', 
+    updated_fields.new_fields
+)
+FROM (
+    SELECT 
+        r_inner.id,
+        jsonb_object_agg(
+            CASE 
+                -- Case 1: childrenCount_NODEUUID_NODEDEFUUID
+                WHEN fields.key LIKE 'childrenCount_%' THEN 
+                    'childrenCount_' || COALESCE(n.i_id::text, split_part(fields.key, '_', 2)) || '_' || split_part(fields.key, '_', 3)
+
+                -- Case 2: Plain NODEUUID
+                ELSE 
+                    COALESCE(n.i_id::text, fields.key)
+            END,
+            fields.value
+        ) as new_fields
+    FROM record r_inner
+    CROSS JOIN LATERAL jsonb_each(r_inner.validation->'fields') AS fields(key, value)
+    -- Join logic: if prefixed, get 2nd part; if not, get 1st part
+    LEFT JOIN node n ON (
+        CASE 
+            WHEN fields.key LIKE 'childrenCount_%' THEN split_part(fields.key, '_', 2)
+            ELSE fields.key
+        END
+    )::uuid = n.uuid
+    GROUP BY r_inner.id
+) AS updated_fields
+WHERE r.id = updated_fields.id;
+
+-- Create an index on the combination of record_uuid and i_id for faster lookups
+
+CREATE INDEX IF NOT EXISTS node_record_uuid_i_id_idx
+	ON node (record_uuid, i_id);
+
+-- Finally, update the node table constraints to reflect the new primary key and foreign key relationships
+
+ALTER TABLE node
+	DROP CONSTRAINT IF EXISTS node_parent_fk,
+	DROP CONSTRAINT IF EXISTS node_pkey,
+	-- Add a composite primary key on record_uuid and i_id
+	ADD PRIMARY KEY (record_uuid, i_id),
+	ADD CONSTRAINT node_parent_fk FOREIGN KEY (record_uuid, p_i_id) REFERENCES "node" (record_uuid, i_id) ON DELETE CASCADE;
+