dimitris.pierrakos
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

added open citations

This commit is contained in:
Antonis Lempesis 2022-04-20 14:46:55 +03:00
parent c442c91f89
commit b7cd2c6ca1
1 changed files with 31 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step13.sql
View File

@ -80,4 +80,34 @@ where reltype='resultResult'
and r2.datainfo.deletedbyinference=false and r2.datainfo.invisible = FALSE and r2.datainfo.deletedbyinference=false and r2.datainfo.invisible = FALSE
and r1.resulttype.classname != 'other' and r1.resulttype.classname != 'other'
and r2.resulttype.classname != 'other' and r2.resulttype.classname != 'other'
and rel.datainfo.deletedbyinference=false and rel.datainfo.invisible = FALSE; and rel.datainfo.deletedbyinference=false and rel.datainfo.invisible = FALSE;
create table ${stats_db_name}.result_citations_oc stored as parquet as
select substr(target, 4) as id, count(distinct substr(source, 4)) as citations
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 relClass='Cites' and rel.datainfo.provenanceaction.classid = 'sysimport:crosswalk:opencitations'
and reltype='resultResult'
and r1.resulttype.classname!=r2.resulttype.classname
and r1.datainfo.deletedbyinference=false and r1.datainfo.invisible = FALSE
and r2.datainfo.deletedbyinference=false and r2.datainfo.invisible = FALSE
and r1.resulttype.classname != 'other'
and r2.resulttype.classname != 'other'
and rel.datainfo.deletedbyinference=false and rel.datainfo.invisible = FALSE
group by substr(target, 4);
create table ${stats_db_name}.result_references_oc stored as parquet as
select substr(source, 4) as id, count(distinct substr(target, 4)) as references
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 relClass='Cites' and rel.datainfo.provenanceaction.classid = 'sysimport:crosswalk:opencitations'
and reltype='resultResult'
and r1.resulttype.classname!=r2.resulttype.classname
and r1.datainfo.deletedbyinference=false and r1.datainfo.invisible = FALSE
and r2.datainfo.deletedbyinference=false and r2.datainfo.invisible = FALSE
and r1.resulttype.classname != 'other'
and r2.resulttype.classname != 'other'
and rel.datainfo.deletedbyinference=false and rel.datainfo.invisible = FALSE
group by substr(source, 4);