antonis.lempesis
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
1
Fork 1
Code Issues Pull Requests Projects Releases Wiki Activity

Changes for result__country in observatory step

This commit is contained in:
dimitrispie 2022-11-03 13:00:01 +02:00
parent 5df9c63963
commit b5e9f73eb1
4 changed files with 37 additions and 46 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/config-default.xml
View File

@ -21,7 +21,7 @@
</property>
<property>
<name>hive_jdbc_url</name>
<value>jdbc:hive2://iis-cdh5-test-m3.ocean.icm.edu.pl:10000</value>
<value>jdbc:hive2://iis-cdh5-test-m3.ocean.icm.edu.pl:10000/;UseNativeQuery=1;?spark.executor.memory=19166291558;spark.yarn.executor.memoryOverhead=3225;spark.driver.memory=11596411699;spark.yarn.driver.memoryOverhead=1228</value>
</property>
<property>
<name>oozie.wf.workflow.notification.url</name>

7
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step13.sql
View File

@ -110,4 +110,9 @@ where relClass='Cites' and rel.datainfo.provenanceaction.classid = 'sysimport:cr
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);
group by substr(source, 4);
CREATE TABLE IF NOT EXISTS ${stats_db_name}.result_country stored as parquet as
SELECT DISTINCT substr(res.id, 4) as id, code as code FROM ${openaire_db_name}.result res
LATERAL VIEW explode(res.country.classid) country as code
WHERE res.datainfo.deletedbyinference = FALSE and res.datainfo.invisible = FALSE;

70
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step21-createObservatoryDB.sql
View File

@ -25,9 +25,8 @@ select
r.type,
c.code as ccode, c.name as cname
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -54,9 +53,8 @@ select
r.type,
r.year
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -83,9 +81,8 @@ select
r.type,
r.year, c.code as ccode, c.name as cname
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -112,9 +109,8 @@ select
r.type,
d.name as dname
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_datasources rd on rd.id=r.id
left outer join ${stats_db_name}.datasource d on d.id=rd.datasource
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
@ -143,9 +139,8 @@ select
r.type,
d.name as dname, c.code as ccode, c.name as cname
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_datasources rd on rd.id=r.id
left outer join ${stats_db_name}.datasource d on d.id=rd.datasource
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
@ -232,9 +227,8 @@ select
r.type,
p.funder as pfunder
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
join ${stats_db_name}.result_projects rp on rp.id=r.id
join ${stats_db_name}.project p on p.id=rp.project
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
@ -263,9 +257,8 @@ select
r.type,
p.funder as pfunder, c.code as ccode, c.name as cname
from ${stats_db_name}.result r
join ${stats_db_name}.result_organization ro on ro.id=r.id
join ${stats_db_name}.organization o on o.id=ro.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
join ${stats_db_name}.result_projects rp on rp.id=r.id
join ${stats_db_name}.project p on p.id=rp.project
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
@ -296,9 +289,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -327,9 +319,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -358,9 +349,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -389,9 +379,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -420,9 +409,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
left outer join ${stats_db_name}.result_pids pids on pids.id=r.id
left outer join ${observatory_db_name}.result_cc_licence rln on rln.id=r.id
@ -513,9 +501,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
join ${stats_db_name}.result_projects rp on rp.id=r.id
join ${stats_db_name}.project p on p.id=rp.project
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id
@ -546,9 +533,8 @@ select
from ${stats_db_name}.result r
join ${stats_db_name}.result_datasources rd on rd.id=r.id
join ${stats_db_name}.datasource d on d.id=rd.datasource and d.type in ('Institutional Repository','Data Repository', 'Repository', 'Publication Repository')
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
join ${stats_db_name}.country c on c.code=o.country and c.continent_name='Europe'
join ${stats_db_name}.result_country rc on rc.id=r.id
join ${stats_db_name}.country c on c.code=rc.code and c.continent_name='Europe'
join ${stats_db_name}.result_projects rp on rp.id=r.id
join ${stats_db_name}.project p on p.id=rp.project
left outer join ${stats_db_name}.result_licenses rl on rl.id=r.id

4
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/workflow.xml
View File

@ -1,4 +1,4 @@
<workflow-app name="Graph Stats" xmlns="uri:oozie:workflow:0.5">
<workflow-app name="Graph Stats testobservatory" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>stats_db_name</name>
@ -70,7 +70,7 @@
</configuration>
</global>
<start to="Step1"/>
<start to="step21-createObservatoryDB-pre"/>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>