From 1566046a8a2982d28d0f68e86932132e6ddc5307 Mon Sep 17 00:00:00 2001
From: Sebastian Nagel <sebastian@commoncrawl.org>
Date: Wed, 1 Oct 2025 15:33:03 +0200
Subject: [PATCH] Bug fix counting top-level domains in node list

For the last block of domain names in reversed notation,
sharing the same TLD prefix, two counts are returned:
the correct count and count "1".
---
 src/main/java/org/commoncrawl/webgraph/explore/Graph.java | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/src/main/java/org/commoncrawl/webgraph/explore/Graph.java b/src/main/java/org/commoncrawl/webgraph/explore/Graph.java
index 1b944e4..6fec31a 100644
--- a/src/main/java/org/commoncrawl/webgraph/explore/Graph.java
+++ b/src/main/java/org/commoncrawl/webgraph/explore/Graph.java
@@ -255,8 +255,11 @@ public Stream<Entry<String, Long>> topLevelDomainCounts(IntStream vertexIds) {
 						}
 						count++;
 					}
-					curr = next;
 					res.add(new SimpleEntry<>(tld, count));
+					curr = next;
+					if (!iter.hasNext()) {
+						break;
+					}
 				} while (curr > -1);
 			}
 			return res.stream().sorted(Collections.reverseOrder(Map.Entry.comparingByValue()));