lsmyrnaios
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

refactoring and added "left" as join type to be 100% sure to get the whole set of projects

This commit is contained in:
Miriam Baglioni 2021-05-21 11:49:05 +02:00
parent e07c3ba089
commit 1ee8f13580
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
dhp-workflows/dhp-aggregation/src/main/java/eu/dnetlib/dhp/actionmanager/project/SparkAtomicActionJob.java
View File

@ -120,7 +120,6 @@ public class SparkAtomicActionJob {
.map((MapFunction<Tuple2<CSVProject, CSVProgramme>, Project>) c -> { .map((MapFunction<Tuple2<CSVProject, CSVProgramme>, Project>) c -> {
CSVProject csvProject = c._1(); CSVProject csvProject = c._1();
Optional<CSVProgramme> ocsvProgramme = Optional.ofNullable(c._2());
return Optional return Optional
.ofNullable(c._2()) .ofNullable(c._2())
@ -135,9 +134,9 @@ public class SparkAtomicActionJob {
H2020Programme pm = new H2020Programme(); H2020Programme pm = new H2020Programme();
H2020Classification h2020classification = new H2020Classification(); H2020Classification h2020classification = new H2020Classification();
pm.setCode(csvProject.getProgramme()); pm.setCode(csvProject.getProgramme());
h2020classification.setClassification(ocsvProgramme.get().getClassification()); h2020classification.setClassification(csvProgramme.getClassification());
h2020classification.setH2020Programme(pm); h2020classification.setH2020Programme(pm);
setLevelsandProgramme(h2020classification, ocsvProgramme.get().getClassification_short()); setLevelsandProgramme(h2020classification, csvProgramme.getClassification_short());
// setProgramme(h2020classification, ocsvProgramme.get().getClassification()); // setProgramme(h2020classification, ocsvProgramme.get().getClassification());
pp.setH2020classification(Arrays.asList(h2020classification)); pp.setH2020classification(Arrays.asList(h2020classification));
@ -145,10 +144,11 @@ public class SparkAtomicActionJob {
}) })
.orElse(null); .orElse(null);
}, Encoders.bean(Project.class)); }, Encoders.bean(Project.class))
.filter(Objects::nonNull);
aaproject aaproject
.joinWith(topic, aaproject.col("h2020topiccode").equalTo(topic.col("code"))) .joinWith(topic, aaproject.col("h2020topiccode").equalTo(topic.col("code")), "left")
.map((MapFunction<Tuple2<Project, EXCELTopic>, Project>) p -> { .map((MapFunction<Tuple2<Project, EXCELTopic>, Project>) p -> {
Optional<EXCELTopic> op = Optional.ofNullable(p._2()); Optional<EXCELTopic> op = Optional.ofNullable(p._2());
Project rp = p._1(); Project rp = p._1();