D-Net
/
dnet-hadoop
13
4
Fork 10
Code Issues 14 Pull Requests 8 Projects 2 Releases 1 Wiki Activity

[stats wf] indicators across stats dbs & updates in the org ids #248

Closed
dimitris.pierrakos wants to merge 1742 commits from beta into beta2master_sept_2022
pull from: beta
merge into: D-Net:beta2master_sept_2022
Conversation 3 Commits 1742 Files Changed 1525 +1 -3
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit ece40adc09 - Show all commits

4
dhp-workflows/dhp-graph-mapper/src/main/java/eu/dnetlib/dhp/oa/graph/clean/country/GetDatasourceFromCountry.java
View File

@ -65,7 +65,6 @@ public class GetDatasourceFromCountry implements Serializable {
conf, conf,
isSparkSessionManaged, isSparkSessionManaged,
spark -> { spark -> {
getDatasourceFromCountry(spark, country, inputPath, workingPath); getDatasourceFromCountry(spark, country, inputPath, workingPath);
}); });
} }
@ -83,7 +82,6 @@ public class GetDatasourceFromCountry implements Serializable {
(FilterFunction<Organization>) o -> !o.getDataInfo().getDeletedbyinference() && (FilterFunction<Organization>) o -> !o.getDataInfo().getDeletedbyinference() &&
o.getCountry().getClassid().length() > 0 && o.getCountry().getClassid().length() > 0 &&
o.getCountry().getClassid().equals(country)); o.getCountry().getClassid().equals(country));
;
// filtering of the relations taking the non deleted by inference and those with IsProvidedBy as relclass // filtering of the relations taking the non deleted by inference and those with IsProvidedBy as relclass
Dataset<Relation> relation = spark Dataset<Relation> relation = spark
@ -97,7 +95,7 @@ public class GetDatasourceFromCountry implements Serializable {
!rel.getDataInfo().getDeletedbyinference()); !rel.getDataInfo().getDeletedbyinference());
organization 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()) .map((MapFunction<Tuple2<Organization, Relation>, String>) t2 -> t2._2().getSource(), Encoders.STRING())
.write() .write()
.mode(SaveMode.Overwrite) .mode(SaveMode.Overwrite)