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

Merge pull request 'Various fixes for the stats DB update workflow, step16-createIndicatorsTables.sql' (#425) from stats_step16_fix into master 

Reviewed-on: #425
This commit is contained in:
Claudio Atzori 2024-04-18 11:25:24 +02:00
parent 0486cea4c4 18fdaaf548
commit 8fdd0244ad
1 changed files with 17 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step16-createIndicatorsTables.sql
View File

@ -237,16 +237,16 @@ create table if not exists ${stats_db_name}.indi_pub_gold_oa stored as parquet a
UNION ALL UNION ALL
select id, issn_online as issn from ${stats_db_name}.datasource d where d.id like '%doajarticles%' select id, issn_online as issn from ${stats_db_name}.datasource d where d.id like '%doajarticles%'
UNION ALL UNION ALL
select id, issn_printed as issn from ${stats_db_name}.datasource d join gold_oa on gold_oa.issn=d.issn_printed select id, issn_printed as issn from ${stats_db_name}.datasource d left semi join gold_oa on gold_oa.issn=d.issn_printed
UNION ALL UNION ALL
select id, issn_online as issn from ${stats_db_name}.datasource d join gold_oa on gold_oa.issn=d.issn_online) foo select id, issn_online as issn from ${stats_db_name}.datasource d left semi join gold_oa on gold_oa.issn=d.issn_online) foo
) )
SELECT DISTINCT pd.id, coalesce(is_gold, 0) as is_gold SELECT DISTINCT pd.id, coalesce(is_gold, 0) as is_gold
FROM ${stats_db_name}.publication_datasources pd FROM ${stats_db_name}.publication_datasources pd
left outer join ( left outer join (
select pd.id, 1 as is_gold select pd.id, 1 as is_gold
FROM ${stats_db_name}.publication_datasources pd FROM ${stats_db_name}.publication_datasources pd
join dd on dd.id=pd.datasource left semi join dd on dd.id=pd.datasource
left outer join ${stats_db_name}.result_accessroute ra on ra.id = pd.id where ra.accessroute = 'gold') tmp on tmp.id=pd.id; /*EOS*/ left outer join ${stats_db_name}.result_accessroute ra on ra.id = pd.id where ra.accessroute = 'gold') tmp on tmp.id=pd.id; /*EOS*/
drop table if exists ${stats_db_name}.indi_pub_hybrid_oa_with_cc purge; /*EOS*/ drop table if exists ${stats_db_name}.indi_pub_hybrid_oa_with_cc purge; /*EOS*/
@ -335,8 +335,8 @@ select ar.organization, rf.no_result_fair/ar.no_allresults org_fairness
from allresults ar from allresults ar
join result_fair rf on rf.organization=ar.organization; /*EOS*/ join result_fair rf on rf.organization=ar.organization; /*EOS*/
DROP VIEW result_fair; DROP VIEW result_fair; /*EOS*/
DROP VIEW allresults; DROP VIEW allresults; /*EOS*/
CREATE TEMPORARY VIEW result_fair as CREATE TEMPORARY VIEW result_fair as
select year, ro.organization organization, count(distinct ro.id) no_result_fair from ${stats_db_name}.result_organization ro select year, ro.organization organization, count(distinct ro.id) no_result_fair from ${stats_db_name}.result_organization ro
@ -1000,13 +1000,18 @@ left outer join (
drop table if exists ${stats_db_name}.result_country purge; /*EOS*/ drop table if exists ${stats_db_name}.result_country purge; /*EOS*/
create table ${stats_db_name}.result_country stored as parquet as create table ${stats_db_name}.result_country stored as parquet as
select distinct ro.id, coalesce(o.country, f.country) select distinct *
from ${stats_db_name}.result_organization ro from (
left outer join ${stats_db_name}.organization o on o.id=ro.organization select ro.id, o.country
left outer join ${stats_db_name}.result_projects rp on rp.id=ro.id from ${stats_db_name}.result_organization ro
left outer join ${stats_db_name}.project p on p.id=rp.project left outer join ${stats_db_name}.organization o on o.id=ro.organization
left outer join ${stats_db_name}.funder f on f.name=p.funder union all
where coalesce(o.country, f.country) IS NOT NULL; select rp.id, f.country
from ${stats_db_name}.result_projects
left outer join ${stats_db_name}.project p on p.id=rp.project
left outer join ${stats_db_name}.funder f on f.name=p.funder
) rc
where rc.country is not null; /*EOS*/
drop table if exists ${stats_db_name}.indi_result_oa_with_license purge; /*EOS*/ drop table if exists ${stats_db_name}.indi_result_oa_with_license purge; /*EOS*/
create table ${stats_db_name}.indi_result_oa_with_license stored as parquet as create table ${stats_db_name}.indi_result_oa_with_license stored as parquet as