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

Update step16-createIndicatorsTables.sql 

- Add org names to indi_project_collab_org
- Add indi_pub_bronze_oa
 - Changes to indi_pub_hybrid_oa_with_cc
This commit is contained in:
dimitrispie 2023-05-25 14:52:34 +03:00
parent 86f4f63daf
commit d6102dd576
1 changed files with 45 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

51
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step16-createIndicatorsTables.sql
View File

@ -119,12 +119,16 @@ drop table tmp purge;
ANALYZE TABLE indi_result_org_country_collab COMPUTE STATISTICS;
create TEMPORARY TABLE AS
select o.id organization, o.name, ro.project as project from organization o
join organization_projects ro on o.id=ro.id;
create table if not exists indi_project_collab_org stored as parquet as
select o1.id org1,o2.id org2, count(distinct o1.project) as collaborations
from organization_projects as o1
join organization_projects as o2 on o1.project=o2.project
where o1.id!=o2.id
group by o1.id, o2.id;
select o1.organization org1,o1.name orgname1, o2.organization org2, o2.name orgname2, count(distinct o1.project) as collaborations
from tmp as o1
join tmp as o2 on o1.project=o2.project
where o1.organization<>o2.organization and o1.name<>o2.name
group by o1.name,o2.name, o1.organization, o2.organization;
ANALYZE TABLE indi_project_collab_org COMPUTE STATISTICS;
@ -245,10 +249,45 @@ FROM publication_datasources pd
JOIN issn on issn.id=pd.datasource
JOIN hybrid_oa ON issn.issn = hybrid_oa.issn
JOIN indi_result_has_cc_licence cc on pd.id=cc.id
where cc.has_cc_license=1) tmp on pd.id=tmp.id;
JOIN indi_pub_gold_oa ga on pd.id=ga.id
where cc.has_cc_license=1 and ga.is_gold=0) tmp on pd.id=tmp.id;
ANALYZE TABLE indi_pub_hybrid_oa_with_cc COMPUTE STATISTICS;
create table if not exists indi_pub_bronze_oa stored as parquet as
WITH hybrid_oa AS (
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_print as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_print != ""
UNION ALL
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_online as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_online != "" and (journal_is_in_doaj = FALSE OR journal_is_oa = FALSE)),
issn AS (
SELECT *
FROM (
SELECT id, issn_printed as issn
FROM datasource
WHERE issn_printed IS NOT NULL
UNION ALL
SELECT id,issn_online as issn
FROM datasource
WHERE issn_online IS NOT NULL ) as issn
WHERE LENGTH(issn) > 7)
SELECT DISTINCT pd.id, coalesce(is_bronze_oa, 0) as is_hybrid_oa
FROM publication_datasources pd
LEFT OUTER JOIN (
SELECT pd.id, 1 as is_bronze_oa from publication_datasources pd
JOIN datasource d on d.id=pd.datasource
JOIN issn on issn.id=pd.datasource
JOIN hybrid_oa ON issn.issn = hybrid_oa.issn
JOIN indi_result_has_cc_licence cc on pd.id=cc.id
JOIN indi_pub_gold_oa ga on pd.id=ga.id
JOIN indi_pub_hybrid_oa_with_cc hy on hy.id=pd.id
where cc.has_cc_license=0 and ga.is_gold=0 and hy.is_hybrid_oa=0) tmp on pd.id=tmp.id;
ANALYZE TABLE indi_pub_bronze_oa COMPUTE STATISTICS;
create table if not exists indi_pub_downloads stored as parquet as
SELECT result_id, sum(downloads) no_downloads from openaire_prod_usage_stats.usage_stats
join publication on result_id=id