Separated impala from hive steps

This commit is contained in:
Antonis Lempesis 2020-07-02 02:42:30 +03:00 committed by Spyros Zoupanos
parent de49173420
commit e6f50de6ef
5 changed files with 89 additions and 60 deletions

View File

@ -0,0 +1,6 @@
-- datasource sources:
-- where the datasource info have been collected from.
create table if not exists ${stats_db_name}.datasource_sources AS
select substr(d.id,4) as id, substr(cf.key, 4) as datasource
from ${openaire_db_name}.datasource d lateral view explode(d.collectedfrom) cfrom as cf
where d.datainfo.deletedbyinference=false;

View File

@ -2,7 +2,6 @@
-- Shortcuts for various definitions in stats db ---
----------------------------------------------------
-- Peer reviewed:
-- Results that have been collected from Crossref
create table ${stats_db_name}.result_peerreviewed as
@ -75,59 +74,3 @@ join ${stats_db_name}.datasource d on d.id=rd.datasource
join ${stats_db_name}.datasource_organizations dor on dor.id=d.id
join ${stats_db_name}.organization o on o.id=dor.organization
where o.country is not null and o.country!='';
-- replace the creation of the result view to include the boolean fields from the previous tables (green, gold,
-- peer reviewed)
drop table if exists ${stats_db_name}.result_tmp;
CREATE TABLE result_tmp (
id STRING,
title STRING,
publisher STRING,
journal STRING,
`date` STRING,
`year` INT,
bestlicence STRING,
access_mode STRING,
embargo_end_date STRING,
delayed BOOLEAN,
authors INT,
source STRING,
abstract BOOLEAN,
type STRING ,
peer_reviewed BOOLEAN,
green BOOLEAN,
gold BOOLEAN)
clustered by (id) into 100 buckets stored as orc tblproperties('transactional'='true');
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.publication r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.dataset r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.software r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.otherresearchproduct r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
drop table if exists ${stats_db_name}.result;
drop view if exists ${stats_db_name}.result;
create table ${stats_db_name}.result stored as parquet as select * from ${stats_db_name}.result_tmp;
drop table ${stats_db_name}.result_tmp;

View File

@ -0,0 +1,55 @@
-- replace the creation of the result view to include the boolean fields from the previous tables (green, gold,
-- peer reviewed)
drop table if exists ${stats_db_name}.result_tmp;
CREATE TABLE result_tmp (
id STRING,
title STRING,
publisher STRING,
journal STRING,
`date` STRING,
`year` INT,
bestlicence STRING,
access_mode STRING,
embargo_end_date STRING,
delayed BOOLEAN,
authors INT,
source STRING,
abstract BOOLEAN,
type STRING ,
peer_reviewed BOOLEAN,
green BOOLEAN,
gold BOOLEAN)
clustered by (id) into 100 buckets stored as orc tblproperties('transactional'='true');
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.publication r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.dataset r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.software r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
insert into ${stats_db_name}.result_tmp
select r.id, r.title, r.publisher, r.journal, r.`date`, date_format(r.`date`, 'yyyy'), r.bestlicence, r.bestlicence, r.embargo_end_date, r.delayed, r.authors, r.source, r.abstract, r.type, pr.peer_reviewed, green.green, gold.gold
FROM ${stats_db_name}.otherresearchproduct r
LEFT OUTER JOIN ${stats_db_name}.result_peerreviewed pr on pr.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_greenoa green on green.id=r.id
LEFT OUTER JOIN ${stats_db_name}.result_gold gold on gold.id=r.id;
drop table if exists ${stats_db_name}.result;
drop view if exists ${stats_db_name}.result;
create table ${stats_db_name}.result stored as parquet as select * from ${stats_db_name}.result_tmp;
drop table ${stats_db_name}.result_tmp;

View File

@ -4,7 +4,7 @@
----------------------------------------------------------------
----------------------------------------------------------------
DROP TABLE IF EXISTS ${stats_db_name}.organization;
CREATE TABLE IF NOT EXISTS ${stats_db_name}.organization AS SELECT substr(o.id, 4) as id, o.legalname.value as name, o.legalshortname as legalshortname, o.country.classid as country
CREATE TABLE IF NOT EXISTS ${stats_db_name}.organization AS SELECT substr(o.id, 4) as id, o.legalname.value as name, o.legalshortname.value as legalshortname, o.country.classid as country
FROM ${openaire_db_name}.organization o WHERE o.datainfo.deletedbyinference=FALSE;
CREATE OR REPLACE VIEW ${stats_db_name}.organization_datasources AS SELECT organization AS id, id AS datasource FROM ${stats_db_name}.datasource_organizations;

View File

@ -215,14 +215,39 @@
<param>stats_db_name=${stats_db_name}</param>
<param>openaire_db_name=${openaire_db_name}</param>
</hive2>
<ok to="Step15_5"/>
<error to="Kill"/>
</action>
<action name="Step15_5">
<hive2 xmlns="uri:oozie:hive2-action:0.1">
<jdbc-url>${hive_jdbc_url}</jdbc-url>
<script>scripts/step15_5.sql</script>
<param>stats_db_name=${stats_db_name}</param>
<param>openaire_db_name=${openaire_db_name}</param>
</hive2>
<ok to="Step16"/>
<error to="Kill"/>
</action>
<action name="Step16">
<shell xmlns="uri:oozie:shell-action:0.1">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<exec>impala-shell.sh</exec>
<argument>${stats_db_name}</argument>
<argument>step16.sql</argument>
<argument>/user/${wf:user()}/oa/graph/stats/oozie_app/scripts/step16.sql</argument>
<file>impala-shell.sh</file>
</shell>
<ok to="Step16_5"/>
<error to="Kill"/>
</action>
<action name="Step16_5">
<hive2 xmlns="uri:oozie:hive2-action:0.1">
<jdbc-url>${hive_jdbc_url}</jdbc-url>
<script>scripts/step16.sql</script>
<script>scripts/step16_5.sql</script>
<param>stats_db_name=${stats_db_name}</param>
<param>openaire_db_name=${openaire_db_name}</param>
</hive2>