D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

added result_result relations

This commit is contained in:
Antonis Lempesis 2021-12-23 15:46:17 +02:00
parent 7551e52e95
commit 81ee654271
2 changed files with 17 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step13.sql
View File

@ -67,4 +67,17 @@ from (
LATERAL VIEW explode(author) a as auth
LATERAL VIEW explode(auth.pid) ap as auth_pid
LATERAL VIEW explode(auth.pid.qualifier.classid) apt as author_pid_type
WHERE res.datainfo.deletedbyinference = FALSE and res.datainfo.invisible = FALSE and author_pid_type = 'orcid') as res
WHERE res.datainfo.deletedbyinference = FALSE and res.datainfo.invisible = FALSE and author_pid_type = 'orcid') as res;
create table ${stats_db_name}.result_result stored as parquet as
select substr(rel.source, 4) as source, substr(rel.target, 4) as target, relclass, subreltype
from ${openaire_db_name}.relation rel
join ${openaire_db_name}.result r1 on rel.source=r1.id
join ${openaire_db_name}.result r2 on r2.id=rel.target
where reltype='resultResult'
and r1.resulttype.classname!=r2.resulttype.classname
and r1.datainfo.deletedbyinference=false
and r2.datainfo.deletedbyinference=false
and r1.resulttype.classname != 'other'
and r2.resulttype.classname != 'other'
and rel.datainfo.deletedbyinference=false;

3
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step20-createMonitorDB.sql
View File

@ -81,6 +81,9 @@ compute stats TARGET.result_sources;
create table TARGET.result_topics stored as parquet as select * from SOURCE.result_topics orig where exists (select 1 from TARGET.result r where r.id=orig.id);
compute stats TARGET.result_topics;
create table TARGET.result_result stored as parquet as select * from SOURCE.result_result orig where exists (select 1 from TARGET.result r where r.id=orig.source or r.id=orig.target);
compute stats TARGET.result_result;
-- datasources
create view if not exists TARGET.datasource as select * from SOURCE.datasource;
create view if not exists TARGET.datasource_oids as select * from SOURCE.datasource_oids;