when deleting duplicates, always keep the oldest node (preserve history)

Author: k-yle

src/__tests__/mock/linz-dump.csv

@@ -199,3 +199,4 @@ WKT,address_id,change_id,address_class,unit_value,address_number,address_number_
 ,70242,,Road,,8,,,,,Oakleigh,,8,Side Street,,,,,,,,,,,174.242,-36.9
 ,70243,,Road,,243,,,,,Oakleigh,,243,Example Street,,,,,,,,,,,174.243,-36.9
 ,70244,,Road,,9,,,,,Oakleigh,,9,Side Street,,,,,,,,,,,174.244,-36.9
+,70245,,Road,,245,,,,,Oakleigh,,245,Example Street,,,,,,,,,,,174.245,-36.9

src/__tests__/mock/planet.xml

@@ -1148,6 +1148,20 @@
 
   <!-- see ways for test case 243, which is the same as the above, but for a building -->
 
+  <!-- 2 nodes, exact duplicates. 246 is newer (i.e. the node ID is larger), so it should be deleted -->
+  <node id="245" version="1" changeset="1" lat="-36.9" lon="174.245" user="" uid="0" visible="true" timestamp="2012-05-19T11:22:33Z">
+    <tag k="addr:housenumber" v="245" />
+    <tag k="addr:street" v="Example Street" />
+    <tag k="addr:hamlet" v="Oakleigh" />
+    <tag k="ref:linz:address_id" v="70245" />
+  </node>
+  <node id="246" version="1" changeset="1" lat="-36.9" lon="174.246" user="" uid="0" visible="true" timestamp="2012-05-19T11:22:33Z">
+    <tag k="addr:housenumber" v="245" />
+    <tag k="addr:street" v="Example Street" />
+    <tag k="addr:hamlet" v="Oakleigh" />
+    <tag k="ref:linz:address_id" v="70245" />
+  </node>
+
   <!-- <ways> -->
     <way id="31" version="1" changeset="1" user="" uid="0" visible="true" timestamp="2012-05-19T11:22:33Z">
       <nd ref="28" />

src/__tests__/snapshot/duplicate-linz-ref.txt

@@ -7,3 +7,4 @@ Oakleigh
 70137	✅	exists on https://osm.org/node/137 and https://osm.org/way/141
 70147	❌	exists on https://osm.org/node/147 and https://osm.org/node/148
 70149	⚠️	exists on https://osm.org/node/149 and https://osm.org/node/150 and https://osm.org/node/151
+70245	✅	exists on https://osm.org/node/245 and https://osm.org/node/246

src/__tests__/snapshot/index.json

@@ -270,9 +270,9 @@
       "url": "https://osm-nz.github.io/linz-address-import/suburbs/MergeduplicateaddressesAuckland_140dpit.osmPatch.geo.json",
       "name": "Merge duplicate addresses - Auckland",
       "title": "Merge duplicate addresses - Auckland",
-      "totalCount": 3,
+      "totalCount": 4,
       "source": "",
-      "snippet": "delete 2, edit 1",
+      "snippet": "delete 3, edit 1",
       "extent": [
         [
           174.1509,

src/__tests__/snapshot/stats.json

@@ -5,7 +5,7 @@
     "EXISTS_BUT_WRONG_DATA": 44,
     "EXISTS_BUT_NO_LINZ_REF": 5,
     "MULTIPLE_EXIST_BUT_NO_LINZ_REF": 1,
-    "MULTIPLE_EXIST": 4,
+    "MULTIPLE_EXIST": 5,
     "EXISTS_BUT_LOCATION_WRONG": 5,
     "TOTALLY_MISSING": 38,
     "NEEDS_DELETE": 15,
@@ -16,5 +16,5 @@
     "NEEDS_DELETE_ON_BUILDING": 1,
     "REPLACED_BY_BUILDING": 1
   },
-  "total": 166
+  "total": 167
 }

src/__tests__/snapshot/suburbs/MergeduplicateaddressesAuckland_140dpit.osmPatch.geo.json

@@ -103,6 +103,40 @@
       "properties": {
         "__action": "delete"
       }
+    },
+    {
+      "type": "Feature",
+      "id": "n246",
+      "geometry": {
+        "type": "Polygon",
+        "coordinates": [
+          [
+            [
+              174.2461,
+              -36.8999
+            ],
+            [
+              174.2461,
+              -36.90010000000001
+            ],
+            [
+              174.2459,
+              -36.90010000000001
+            ],
+            [
+              174.2459,
+              -36.8999
+            ],
+            [
+              174.2461,
+              -36.8999
+            ]
+          ]
+        ]
+      },
+      "properties": {
+        "__action": "delete"
+      }
     }
   ],
   "bbox": {

src/action/handlers/handleDuplicateLinzRef.ts

@@ -26,9 +26,10 @@ export async function handleDuplicateLinzRef(
   const features: GeoJsonFeature[] = [];
 
   for (const [linzId, [linzAddr, osmAddrList]] of array) {
-    const simpleNodes = osmAddrList.filter(
-      (x) => x.osmId[0] === 'n' && !x.isNonTrivial,
-    );
+    const simpleNodes = osmAddrList
+      .filter((x) => x.osmId[0] === 'n' && !x.isNonTrivial)
+      // if there are duplicates, keep the oldest node (determined by the node ID)
+      .sort((a, b) => +a.osmId.slice(1) - +b.osmId.slice(1));
 
     // we can autofix this if some of the duplicates are simple nodes
 
@@ -55,7 +56,7 @@ export async function handleDuplicateLinzRef(
         autofixable[linzId] = '✅';
       }
     } else if (simpleNodes.length) {
-      // Either (a) all nodes are simple. Arbitrarily pick 1 to keep and delete the rest.
+      // Either (a) all nodes are simple. Pick the oldest 1 to keep and delete the rest.
       // Or     ( b) notall of the addresses are simple, so delete only the simple ones.
       const toDelete =
         simpleNodes.length === osmAddrList.length