diff --git a/dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/SparkCountryPropagationJob.java b/dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/SparkCountryPropagationJob.java
index 56aa953b4..25cd82248 100644
--- a/dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/SparkCountryPropagationJob.java
+++ b/dhp-workflows/dhp-enrichment/src/main/java/eu/dnetlib/dhp/countrypropagation/SparkCountryPropagationJob.java
@@ -102,21 +102,27 @@ public class SparkCountryPropagationJob {
 	private static <R extends Result> MapFunction<Tuple2<R, ResultCountrySet>, R> getCountryMergeFn() {
 		return t -> {
 			Optional.ofNullable(t._2()).ifPresent(r -> {
-				t._1().getCountry().addAll(merge(t._1().getCountry(), r.getCountrySet()));
+				if(Optional.ofNullable(t._1().getCountry()).isPresent())
+					t._1().getCountry().addAll(merge(t._1().getCountry(), r.getCountrySet()));
+				else
+					t._1().setCountry(merge(null, t._2().getCountrySet()));
 			});
 			return t._1();
 		};
 	}
 
 	private static List<Country> merge(List<Country> c1, List<CountrySbs> c2) {
-		HashSet<String> countries = c1
-			.stream()
-			.map(Qualifier::getClassid)
-			.collect(Collectors.toCollection(HashSet::new));
+		HashSet<String> countries = new HashSet<>();
+				if(Optional.ofNullable(c1).isPresent()){
+					countries = c1.stream().map(Qualifier::getClassid)
+							.collect(Collectors.toCollection(HashSet::new));
+				}
 
+
+		HashSet<String> finalCountries = countries;
 		return c2
 			.stream()
-			.filter(c -> !countries.contains(c.getClassid()))
+			.filter(c -> !finalCountries.contains(c.getClassid()))
 			.map(c -> getCountry(c.getClassid(), c.getClassname()))
 			.collect(Collectors.toList());
 	}