sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

all indicator tables are now stored as parquet

This commit is contained in:
Antonis Lempesis 2024-04-19 12:45:36 +03:00
parent 0c71c58df6
commit c3fe9662b2
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step16-createIndicatorsTables.sql
View File

@ -380,7 +380,7 @@ CREATE TEMPORARY VIEW allresults as
drop table if exists ${stats_db_name}.indi_org_fairness_pub purge; /*EOS*/ drop table if exists ${stats_db_name}.indi_org_fairness_pub purge; /*EOS*/
create table if not exists ${stats_db_name}.indi_org_fairness_pub as create table if not exists ${stats_db_name}.indi_org_fairness_pub stored as parquet as
select ar.organization, rf.no_result_fair/ar.no_allresults org_fairness select ar.organization, rf.no_result_fair/ar.no_allresults org_fairness
from allresults ar join result_fair rf from allresults ar join result_fair rf
on rf.organization=ar.organization; /*EOS*/ on rf.organization=ar.organization; /*EOS*/
@ -639,7 +639,7 @@ from ${stats_db_name}.publication p
drop table if exists ${stats_db_name}.indi_result_with_pid purge; /*EOS*/ drop table if exists ${stats_db_name}.indi_result_with_pid purge; /*EOS*/
create table if not exists ${stats_db_name}.indi_result_with_pid as create table if not exists ${stats_db_name}.indi_result_with_pid stored as parquet as
select distinct p.id, coalesce(result_with_pid, 0) as result_with_pid select distinct p.id, coalesce(result_with_pid, 0) as result_with_pid
from ${stats_db_name}.result p from ${stats_db_name}.result p
left outer join ( left outer join (
@ -653,7 +653,7 @@ group by rf.id; /*EOS*/
drop table if exists ${stats_db_name}.indi_pub_interdisciplinarity purge; /*EOS*/ drop table if exists ${stats_db_name}.indi_pub_interdisciplinarity purge; /*EOS*/
create table if not exists ${stats_db_name}.indi_pub_interdisciplinarity as create table if not exists ${stats_db_name}.indi_pub_interdisciplinarity stored as parquet as
select distinct p.id as id, coalesce(is_interdisciplinary, 0) select distinct p.id as id, coalesce(is_interdisciplinary, 0)
as is_interdisciplinary as is_interdisciplinary
from pub_fos_totals p from pub_fos_totals p