D-Net
/
dhp-graph-dump
4
0
Fork 0
Code Issues Pull Requests 1 Projects Releases Wiki Activity

[SKG-IFv1.1] Relaxing constraint to include also EDP and SH results

This commit is contained in:
Miriam Baglioni 2024-09-22 18:03:09 +02:00
parent fe71931002
commit 09e2ac1fce
1 changed files with 6 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
dump/src/main/java/eu/dnetlib/dhp/oa/graph/dump/filterentities/SelectLOT1Entities.java
View File

@ -98,8 +98,9 @@ public class SelectLOT1Entities implements Serializable {
.schema(Encoders.bean(Software.class).schema())
.json(inputPath + "software")
.where("datainfo.deletedbyinference != true AND datainfo.invisible != true")
.selectExpr("id", "instance", "explode(pid) as pid")
.where("pid.qualifier.classid IN ('doi', 'swhid')") // filter by pid type
.select("id", "instance", "pid")
.where("array_contains(instance.hostedby.key ,'10|re3data_____::c4b2081b224be6b3e79d0e5e5556f631') OR array_contains(instance.hostedby.key, '10|openaire____::dbfd07503aaa1ed31beed7dec942f3f4') " +
"OR array_contains(pid.qualifier.classid,'doi') OR array_contains(pid.qualifier.classid,'swhid')") // filter by pid type and hosting datasource
.selectExpr("id", "explode(instance) as instance")
.withColumn(
"CCL",
@ -119,8 +120,9 @@ public class SelectLOT1Entities implements Serializable {
.schema(Encoders.bean(Dataset.class).schema())
.json(inputPath + "dataset")
.where("datainfo.deletedbyinference != true AND datainfo.invisible != true")
.selectExpr("id", "instance", "explode(pid) as pid")
.where("pid.qualifier.classid IN ('doi', 'handle', 'pdb', 'ena', 'uniprot')") // filter by pid type
.select("id", "instance", "pid")
.where("array_contains(pid.qualifier.classid ,'doi') OR array_contains(pid.qualifier.classid ,'handle') OR array_contains(pid.qualifier.classid , 'pdb') array_contains(pid.qualifier.classid , 'ena') array_contains(pid.qualifier.classid , 'uniprot') OR " +
"array_contains(instance.hostedby.key ,'10|re3data_____::c4b2081b224be6b3e79d0e5e5556f631') OR array_contains(instance.hostedby.key, '10|openaire____::dbfd07503aaa1ed31beed7dec942f3f4') )") // filter by pid type
.selectExpr("id", "explode(instance) as instance")
.withColumn(
"CCL",