From a1861b9eaa7570c890877ec8e142a49bbebfb12b Mon Sep 17 00:00:00 2001
From: Enrico Ottonello <enrico.ottonello@isti.cnr.it>
Date: Fri, 24 Apr 2020 18:33:37 +0200
Subject: [PATCH] workflow works in parallel on 2 activity files

---
 .../orcid/OrcidAuthorsDOIsDataGen.java        |  6 +---
 .../oozie_app/workflow.xml                    | 32 ++++++++++++++++---
 2 files changed, 28 insertions(+), 10 deletions(-)

diff --git a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/OrcidAuthorsDOIsDataGen.java b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/OrcidAuthorsDOIsDataGen.java
index 7596cf67f..c7c1d75cf 100644
--- a/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/OrcidAuthorsDOIsDataGen.java
+++ b/dhp-workflows/dhp-doiboost/src/main/java/eu/dnetlib/doiboost/orcid/OrcidAuthorsDOIsDataGen.java
@@ -25,11 +25,7 @@ public class OrcidAuthorsDOIsDataGen extends OrcidDSManager {
         String tarGzUri =
                 hdfsServerUri.concat(hdfsOrcidDefaultPath).concat(activitiesFileNameTarGz);
         Path outputPath =
-                new Path(
-                        hdfsServerUri
-                                .concat(hdfsOrcidDefaultPath)
-                                .concat(outputAuthorsDOIsPath)
-                                .concat("authors_dois.seq"));
+                new Path(hdfsServerUri.concat(hdfsOrcidDefaultPath).concat(outputAuthorsDOIsPath));
         ActivitiesDecompressor.parseGzActivities(conf, tarGzUri, outputPath);
     }
 
diff --git a/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_authors_dois_data/oozie_app/workflow.xml b/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_authors_dois_data/oozie_app/workflow.xml
index 5d7222d07..3dabb765b 100644
--- a/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_authors_dois_data/oozie_app/workflow.xml
+++ b/dhp-workflows/dhp-doiboost/src/main/resources/eu/dnetlib/dhp/doiboost/orcid_authors_dois_data/oozie_app/workflow.xml
@@ -18,11 +18,16 @@
             <delete path='${workingPath_activities}/output'/>
             <mkdir path='${workingPath_activities}/output'/>
         </fs>
-        <ok to="GenerateOrcidAuthorsDOIsData"/>
+        <ok to="fork_generate_orcid_authors_dois"/>
         <error to="Kill"/>
     </action>
     
-    <action name="GenerateOrcidAuthorsDOIsData">
+    <fork name = "fork_generate_orcid_authors_dois">
+      <path start = "Generate_Orcid_Authors_DOIs_0"/>
+      <path start = "Generate_Orcid_Authors_DOIs_1"/>
+   	</fork>
+   	
+    <action name="Generate_Orcid_Authors_DOIs_0">
         <java>
             <job-tracker>${jobTracker}</job-tracker>
             <name-node>${nameNode}</name-node>
@@ -30,10 +35,27 @@
             <arg>-d</arg><arg>${workingPath_activities}/</arg>
             <arg>-n</arg><arg>${nameNode}</arg>
             <arg>-f</arg><arg>ORCID_2019_activites_0.tar.gz</arg>
-            <arg>-o</arg><arg>output/</arg>
+            <arg>-o</arg><arg>output/authors_dois_0.seq</arg>
         </java>
-        <ok to="End"/>
+        <ok to="join_node"/>
         <error to="Kill"/>
     </action>
-    <end name="End"/>
+    
+     <action name="Generate_Orcid_Authors_DOIs_1">
+        <java>
+            <job-tracker>${jobTracker}</job-tracker>
+            <name-node>${nameNode}</name-node>
+            <main-class>eu.dnetlib.doiboost.orcid.OrcidAuthorsDOIsDataGen</main-class>
+            <arg>-d</arg><arg>${workingPath_activities}/</arg>
+            <arg>-n</arg><arg>${nameNode}</arg>
+            <arg>-f</arg><arg>ORCID_2019_activites_1.tar.gz</arg>
+            <arg>-o</arg><arg>output/authors_dois_1.seq</arg>
+        </java>
+        <ok to="join_node"/>
+        <error to="Kill"/>
+    </action>
+    
+   <join name = "join_node" to = "End"/>
+    
+   <end name="End"/>
 </workflow-app>
\ No newline at end of file