From 8d97949316a93884a123746cec0d60f6c34b2a2e Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Fri, 7 Oct 2022 09:52:45 +0200
Subject: [PATCH 1/2] [cleaning] fixed loop in wf nodes

---
 .../eu/dnetlib/dhp/oa/graph/clean/oozie_app/workflow.xml    | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/clean/oozie_app/workflow.xml b/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/clean/oozie_app/workflow.xml
index 2ba0a7ad7..08e74a5e5 100644
--- a/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/clean/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/clean/oozie_app/workflow.xml
@@ -492,7 +492,7 @@
             <arg>--datasourcePath</arg><arg>${workingDir}/working/hostedby</arg>
             <arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
         </spark>
-        <ok to="wait_clean_context"/>
+        <ok to="wait_clean_country"/>
         <error to="Kill"/>
     </action>
 
@@ -521,7 +521,7 @@
             <arg>--datasourcePath</arg><arg>${workingDir}/working/hostedby</arg>
             <arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
         </spark>
-        <ok to="wait_clean_context"/>
+        <ok to="wait_clean_country"/>
         <error to="Kill"/>
     </action>
 
@@ -550,7 +550,7 @@
             <arg>--datasourcePath</arg><arg>${workingDir}/working/hostedby</arg>
             <arg>--collectedfrom</arg><arg>${collectedfrom}</arg>
         </spark>
-        <ok to="wait_clean_context"/>
+        <ok to="wait_clean_country"/>
         <error to="Kill"/>
     </action>
 

From ece40adc09b5124386fb4f28391f7b151ea0ceea Mon Sep 17 00:00:00 2001
From: Claudio Atzori <claudio.atzori@isti.cnr.it>
Date: Tue, 11 Oct 2022 10:10:20 +0200
Subject: [PATCH 2/2] [cleaning] fixing NPE in the country cleaning phase

---
 .../dhp/oa/graph/clean/country/GetDatasourceFromCountry.java  | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/country/GetDatasourceFromCountry.java b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/country/GetDatasourceFromCountry.java
index dd5af6998..d3741d3e8 100644
--- a/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/country/GetDatasourceFromCountry.java
+++ b/dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/country/GetDatasourceFromCountry.java
@@ -65,7 +65,6 @@ public class GetDatasourceFromCountry implements Serializable {
 			conf,
 			isSparkSessionManaged,
 			spark -> {
-
 				getDatasourceFromCountry(spark, country, inputPath, workingPath);
 			});
 	}
@@ -83,7 +82,6 @@ public class GetDatasourceFromCountry implements Serializable {
 				(FilterFunction<Organization>) o -> !o.getDataInfo().getDeletedbyinference() &&
 					o.getCountry().getClassid().length() > 0 &&
 					o.getCountry().getClassid().equals(country));
-		;
 
 		// filtering of the relations taking the non deleted by inference and those with IsProvidedBy as relclass
 		Dataset<Relation> relation = spark
@@ -97,7 +95,7 @@ public class GetDatasourceFromCountry implements Serializable {
 					!rel.getDataInfo().getDeletedbyinference());
 
 		organization
-			.joinWith(relation, organization.col("id").equalTo(relation.col("target")), "left")
+			.joinWith(relation, organization.col("id").equalTo(relation.col("target")))
 			.map((MapFunction<Tuple2<Organization, Relation>, String>) t2 -> t2._2().getSource(), Encoders.STRING())
 			.write()
 			.mode(SaveMode.Overwrite)