D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 12 Projects 2 Releases 2 Wiki Activity

Update createIndicatorsTables.sql

This commit is contained in:
dimitrispie 2021-06-22 09:26:49 +03:00
parent da1f123d7b
commit 2cd0514397
1 changed files with 43 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

43
dhp-workflows/dhp-indicators/src/main/resources/eu/dnetlib/dhp/oa/graph/indicators/oozie_app/scripts/createIndicatorsTables.sql
View File

@ -3,5 +3,48 @@ select f.id as id, count(pr.result) as total_pubs from SOURCE.funder f
join SOURCE.project p on f.name=p.funder
join SOURCE.project_results_publication pr on pr.project_results=p.id group by f.id, f.name;
create table TARGET.indi_pub_green_oa stored as parquet as
select distinct p.id, coalesce(green_oa, 0) as green_oa
from SOURCE.publication p
left outer join (
select p.id, 1 as green_oa
from SOURCE.publication p
join SOURCE.result_instance ri on ri.id = p.id
join SOURCE.datasource on datasource.id = ri.hostedby
where SOURCE.datasource.type like '%Repository%'
and (ri.accessright = 'Open Access'
or ri.accessright = 'Embargo')) tmp
on p.id= tmp.id;
create table TARGET.indi_pub_grey_lit stored as parquet as
select distinct p.id, coalesce(grey_lit, 0) as grey_lit
from SOURCE.publication p
left outer join (
select p.id, 1 as grey_lit
from SOURCE.publication p
join SOURCE.result_classifications rt on rt.id = p.id
where rt.type not in ('Article','Part of book or chapter of book','Book','Doctoral thesis','Master thesis','Data Paper', 'Thesis', 'Bachelor thesis', 'Conference object') and
not exists (select 1 from result_classifications rc where type ='Other literature type' and rc.id=p.id)) tmp on p.id=tmp.id;
create table TARGET.indi_pub_doi_from_crossref stored as parquet as
select distinct p.id, coalesce(doi_from_crossref, 0) as doi_from_crossref
from SOURCE.publication p
left outer join
(select ri.id, 1 as doi_from_crossref from SOURCE.result_instance ri
join SOURCE.datasource d on d.id = ri.collectedfrom
where pidtype='Digital Object Identifier' and d.name ='Crossref') tmp
on tmp.id=p.id;
create table TARGET.indi_pub_gold_oa stored as parquet as
select distinct p.id, coalesce(gold_oa, 0) as gold_oa
from SOURCE.publication p
left outer join (
select p.id, 1 as gold_oa
from SOURCE.publication p
join SOURCE.result_instance ri on ri.id = p.id
join SOURCE.datasource on datasource.id = ri.hostedby
where SOURCE.datasource.id like '%doajarticles%') tmp
on p.id= tmp.id;
compute stats TARGET.funders_publications;