D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 10 Projects 2 Releases 2 Wiki Activity

removed module dhp-stats-monitor-update

This commit is contained in:
Claudio Atzori 2023-10-10 09:52:03 +02:00
parent 204404b0e3
commit ed9282ef2a
8 changed files with 0 additions and 598 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/config-default.xml
View File

@ -1,30 +0,0 @@
<configuration>
<property>
<name>jobTracker</name>
<value>${jobTracker}</value>
</property>
<property>
<name>nameNode</name>
<value>${nameNode}</value>
</property>
<property>
<name>oozie.use.system.libpath</name>
<value>true</value>
</property>
<property>
<name>oozie.action.sharelib.for.spark</name>
<value>spark2</value>
</property>
<property>
<name>hive_metastore_uris</name>
<value>thrift://iis-cdh5-test-m3.ocean.icm.edu.pl:9083</value>
</property>
<property>
<name>hive_jdbc_url</name>
<value>jdbc:hive2://iis-cdh5-test-m3.ocean.icm.edu.pl:10000/;UseNativeQuery=1;?spark.executor.memory=22166291558;spark.yarn.executor.memoryOverhead=3225;spark.driver.memory=15596411699;spark.yarn.driver.memoryOverhead=1228</value>
</property>
<property>
<name>oozie.wf.workflow.notification.url</name>
<value>{serviceUrl}/v1/oozieNotification/jobUpdate?jobId=$jobId%26status=$status</value>
</property>
</configuration>

75
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/copyDataToImpalaCluster.sh
View File

@ -1,75 +0,0 @@
export PYTHON_EGG_CACHE=/home/$(whoami)/.python-eggs
export link_folder=/tmp/impala-shell-python-egg-cache-$(whoami)
if ! [ -L $link_folder ]
then
rm -Rf "$link_folder"
ln -sfn ${PYTHON_EGG_CACHE}${link_folder} ${link_folder}
fi
#export HADOOP_USER_NAME=$2
function copydb() {
export HADOOP_USER="dimitris.pierrakos"
export HADOOP_USER_NAME='dimitris.pierrakos'
db=$1
FILE=("hive_wf_tmp_"$RANDOM)
hdfs dfs -mkdir hdfs://impala-cluster-mn1.openaire.eu:8020/tmp/$FILE/
# change ownership to impala
# hdfs dfs -conf /etc/impala_cluster/hdfs-site.xml -chmod -R 777 /tmp/$FILE/${db}.db
hdfs dfs -conf /etc/impala_cluster/hdfs-site.xml -chmod -R 777 /tmp/$FILE/
# copy the databases from ocean to impala
echo "copying $db"
hadoop distcp -Dmapreduce.map.memory.mb=6144 -pb hdfs://nameservice1/user/hive/warehouse/${db}.db hdfs://impala-cluster-mn1.openaire.eu:8020/tmp/$FILE/
hdfs dfs -conf /etc/impala_cluster/hdfs-site.xml -chmod -R 777 /tmp/$FILE/${db}.db
# drop tables from db
for i in `impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} --delimited -q "show tables"`;
do
`impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} -q "drop table $i;"`;
done
# drop views from db
for i in `impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} --delimited -q "show tables"`;
do
`impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} -q "drop view $i;"`;
done
# delete the database
impala-shell -i impala-cluster-dn1.openaire.eu -q "drop database if exists ${db} cascade";
# create the databases
impala-shell -i impala-cluster-dn1.openaire.eu -q "create database ${db}";
impala-shell -q "INVALIDATE METADATA"
echo "creating schema for ${db}"
for (( k = 0; k < 5; k ++ )); do
for i in `impala-shell -d ${db} --delimited -q "show tables"`;
do
impala-shell -d ${db} --delimited -q "show create table $i";
done | sed 's/"$/;/' | sed 's/^"//' | sed 's/[[:space:]]\date[[:space:]]/`date`/g' | impala-shell --user $HADOOP_USER_NAME -i impala-cluster-dn1.openaire.eu -c -f -
done
# load the data from /tmp in the respective tables
echo "copying data in tables and computing stats"
for i in `impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} --delimited -q "show tables"`;
do
impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} -q "load data inpath '/tmp/$FILE/${db}.db/$i' into table $i";
impala-shell -i impala-cluster-dn1.openaire.eu -d ${db} -q "compute stats $i";
done
# deleting the remaining directory from hdfs
hdfs dfs -conf /etc/impala_cluster/hdfs-site.xml -rm -R /tmp/$FILE/${db}.db
}
MONITOR_DB=$1
#HADOOP_USER_NAME=$2
copydb $MONITOR_DB'_institutions'
copydb $MONITOR_DB

29
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/finalizeImpalaCluster.sh
View File

@ -1,29 +0,0 @@
export PYTHON_EGG_CACHE=/home/$(whoami)/.python-eggs
export link_folder=/tmp/impala-shell-python-egg-cache-$(whoami)
if ! [ -L $link_folder ]
then
rm -Rf "$link_folder"
ln -sfn ${PYTHON_EGG_CACHE}${link_folder} ${link_folder}
fi
function createShadowDB() {
SOURCE=$1
SHADOW=$2
# drop views from db
for i in `impala-shell -i impala-cluster-dn1.openaire.eu -d ${SHADOW} --delimited -q "show tables"`;
do
`impala-shell -i impala-cluster-dn1.openaire.eu -d ${SHADOW} -q "drop view $i;"`;
done
impala-shell -i impala-cluster-dn1.openaire.eu -q "drop database ${SHADOW} CASCADE";
impala-shell -i impala-cluster-dn1.openaire.eu -q "create database if not exists ${SHADOW}";
# impala-shell -i impala-cluster-dn1.openaire.eu -d ${SHADOW} -q "show tables" | sed "s/^/drop view if exists ${SHADOW}./" | sed "s/$/;/" | impala-shell -i impala-cluster-dn1.openaire.eu -f -
impala-shell -i impala-cluster-dn1.openaire.eu -d ${SOURCE} -q "show tables" --delimited | sed "s/\(.*\)/create view ${SHADOW}.\1 as select * from ${SOURCE}.\1;/" | impala-shell -i impala-cluster-dn1.openaire.eu -f -
}
MONITOR_DB=$1
MONITOR_DB_SHADOW=$2
createShadowDB $MONITOR_DB'_institutions' $MONITOR_DB'_institutions_shadow'
createShadowDB $MONITOR_DB $MONITOR_DB'_shadow'

54
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/monitor.sh
View File

@ -1,54 +0,0 @@
export PYTHON_EGG_CACHE=/home/$(whoami)/.python-eggs
export link_folder=/tmp/impala-shell-python-egg-cache-$(whoami)
if ! [ -L $link_folder ]
then
rm -Rf "$link_folder"
ln -sfn ${PYTHON_EGG_CACHE}${link_folder} ${link_folder}
fi
export SOURCE=$1
export TARGET=$2
export SHADOW=$3
export SCRIPT_PATH=$4
export SCRIPT_PATH2=$5
export SCRIPT_PATH2=$6
export HIVE_OPTS="-hiveconf mapred.job.queue.name=analytics -hiveconf hive.spark.client.connect.timeout=120000ms -hiveconf hive.spark.client.server.connect.timeout=300000ms -hiveconf spark.executor.memory=19166291558 -hiveconf spark.yarn.executor.memoryOverhead=3225 -hiveconf spark.driver.memory=11596411699 -hiveconf spark.yarn.driver.memoryOverhead=1228"
export HADOOP_USER_NAME="oozie"
echo "Getting file from " $4
hdfs dfs -copyToLocal $4
echo "Getting file from " $5
hdfs dfs -copyToLocal $5
echo "Getting file from " $6
hdfs dfs -copyToLocal $6
#update Institutions DB
cat updateMonitorDB_institutions.sql | sed "s/SOURCE/$1/g" | sed "s/TARGET/$2_institutions/g1" > foo
hive $HIVE_OPTS -f foo
cat updateMonitorDB.sql | sed "s/SOURCE/$1/g" | sed "s/TARGET/$2_institutions/g1" > foo
hive $HIVE_OPTS -f foo
echo "Hive shell finished"
echo "Updating shadow monitor insitutions database"
hive -e "drop database if exists ${SHADOW}_institutions cascade"
hive -e "create database if not exists ${SHADOW}_institutions"
hive $HIVE_OPTS --database ${2}_institutions -e "show tables" | grep -v WARN | sed "s/\(.*\)/create view ${SHADOW}_institutions.\1 as select * from ${2}_institutions.\1;/" > foo
hive -f foo
echo "Shadow db monitor insitutions ready!"
#update Monitor DB
cat updateMonitorDBAll.sql | sed "s/SOURCE/$1/g" | sed "s/TARGET/$2/g1" > foo
hive $HIVE_OPTS -f foo
echo "Hive shell finished"
echo "Updating shadow monitor database"
hive -e "drop database if exists ${SHADOW} cascade"
hive -e "create database if not exists ${SHADOW}"
hive $HIVE_OPTS --database ${2} -e "show tables" | grep -v WARN | sed "s/\(.*\)/create view ${SHADOW}.\1 as select * from ${2}.\1;/" > foo
hive -f foo
echo "Shadow db monitor insitutions ready!"

138
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/scripts/updateMonitorDB.sql
View File

@ -1,138 +0,0 @@
INSERT INTO TARGET.result select * from TARGET.result_new;
ANALYZE TABLE TARGET.result COMPUTE STATISTICS;
INSERT INTO TARGET.result_citations select * from SOURCE.result_citations orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_citations COMPUTE STATISTICS;
INSERT INTO TARGET.result_references_oc select * from SOURCE.result_references_oc orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_references_oc COMPUTE STATISTICS;
INSERT INTO TARGET.result_classifications select * from SOURCE.result_classifications orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_classifications COMPUTE STATISTICS;
INSERT INTO TARGET.result_apc select * from SOURCE.result_apc orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_apc COMPUTE STATISTICS;
INSERT INTO TARGET.result_concepts select * from SOURCE.result_concepts orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_concepts COMPUTE STATISTICS;
INSERT INTO TARGET.result_datasources select * from SOURCE.result_datasources orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_datasources COMPUTE STATISTICS;
INSERT INTO TARGET.result_fundercount select * from SOURCE.result_fundercount orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_fundercount COMPUTE STATISTICS;
INSERT INTO TARGET.result_gold select * from SOURCE.result_gold orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_gold COMPUTE STATISTICS;
INSERT INTO TARGET.result_greenoa select * from SOURCE.result_greenoa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_greenoa COMPUTE STATISTICS;
INSERT INTO TARGET.result_languages select * from SOURCE.result_languages orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_languages COMPUTE STATISTICS;
INSERT INTO TARGET.result_licenses select * from SOURCE.result_licenses orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_licenses COMPUTE STATISTICS;
INSERT INTO TARGET.result_oids select * from SOURCE.result_oids orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_oids COMPUTE STATISTICS;
INSERT INTO TARGET.result_organization select * from SOURCE.result_organization orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_organization COMPUTE STATISTICS;
INSERT INTO TARGET.result_peerreviewed select * from SOURCE.result_peerreviewed orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_peerreviewed COMPUTE STATISTICS;
INSERT INTO TARGET.result_pids select * from SOURCE.result_pids orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_pids COMPUTE STATISTICS;
INSERT INTO TARGET.result_projectcount select * from SOURCE.result_projectcount orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_projectcount COMPUTE STATISTICS;
INSERT INTO TARGET.result_projects select * from SOURCE.result_projects orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_projects COMPUTE STATISTICS;
INSERT INTO TARGET.result_refereed select * from SOURCE.result_refereed orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_refereed COMPUTE STATISTICS;
INSERT INTO TARGET.result_sources select * from SOURCE.result_sources orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_sources COMPUTE STATISTICS;
INSERT INTO TARGET.result_topics select * from SOURCE.result_topics orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_topics COMPUTE STATISTICS;
INSERT INTO TARGET.result_fos select * from SOURCE.result_fos orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_fos COMPUTE STATISTICS;
INSERT INTO TARGET.result_accessroute select * from SOURCE.result_accessroute orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_accessroute COMPUTE STATISTICS;
create or replace view TARGET.foo1 as select * from SOURCE.result_result rr where rr.source in (select id from TARGET.result_new);
create or replace view TARGET.foo2 as select * from SOURCE.result_result rr where rr.target in (select id from TARGET.result_new);
insert into TARGET.result_result select distinct * from (select * from TARGET.foo1 union all select * from TARGET.foo2) foufou;
drop view TARGET.foo1;
drop view TARGET.foo2;
ANALYZE TABLE TARGET.result_result COMPUTE STATISTICS;
-- indicators
-- Sprint 1 ----
INSERT INTO TARGET.indi_pub_green_oa select * from SOURCE.indi_pub_green_oa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_green_oa COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_grey_lit select * from SOURCE.indi_pub_grey_lit orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_grey_lit COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_doi_from_crossref select * from SOURCE.indi_pub_doi_from_crossref orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_doi_from_crossref COMPUTE STATISTICS;
-- Sprint 2 ----
INSERT INTO TARGET.indi_result_has_cc_licence select * from SOURCE.indi_result_has_cc_licence orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_has_cc_licence COMPUTE STATISTICS;
INSERT INTO TARGET.indi_result_has_cc_licence_url select * from SOURCE.indi_result_has_cc_licence_url orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_has_cc_licence_url COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_has_abstract select * from SOURCE.indi_pub_has_abstract orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_has_abstract COMPUTE STATISTICS;
INSERT INTO TARGET.indi_result_with_orcid select * from SOURCE.indi_result_with_orcid orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_with_orcid COMPUTE STATISTICS;
---- Sprint 3 ----
INSERT INTO TARGET.indi_funded_result_with_fundref select * from SOURCE.indi_funded_result_with_fundref orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_funded_result_with_fundref COMPUTE STATISTICS;
---- Sprint 4 ----
INSERT INTO TARGET.indi_pub_diamond select * from SOURCE.indi_pub_diamond orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_diamond COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_in_transformative select * from SOURCE.indi_pub_in_transformative orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_in_transformative COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_closed_other_open select * from SOURCE.indi_pub_closed_other_open orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_closed_other_open COMPUTE STATISTICS;
---- Sprint 5 ----
INSERT INTO TARGET.indi_result_no_of_copies select * from SOURCE.indi_result_no_of_copies orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_no_of_copies COMPUTE STATISTICS;
---- Sprint 6 ----
INSERT INTO TARGET.indi_pub_hybrid_oa_with_cc select * from SOURCE.indi_pub_hybrid_oa_with_cc orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_hybrid_oa_with_cc COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_bronze_oa select * from SOURCE.indi_pub_bronze_oa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_bronze_oa COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads select * from SOURCE.indi_pub_downloads orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads_datasource select * from SOURCE.indi_pub_downloads_datasource orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads_datasource COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads_year select * from SOURCE.indi_pub_downloads_year orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads_year COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads_datasource_year select * from SOURCE.indi_pub_downloads_datasource_year orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads_datasource_year COMPUTE STATISTICS;
---- Sprint 7 ----
INSERT INTO TARGET.indi_pub_gold_oa select * from SOURCE.indi_pub_gold_oa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_gold_oa COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_hybrid select * from SOURCE.indi_pub_hybrid orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_hybrid COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_has_preprint select * from SOURCE.indi_pub_has_preprint orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_has_preprint COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_in_subscribed select * from SOURCE.indi_pub_in_subscribed orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_in_subscribed COMPUTE STATISTICS;
INSERT INTO TARGET.indi_result_with_pid select * from SOURCE.indi_result_with_pid orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_with_pid COMPUTE STATISTICS;
INSERT INTO TARGET.indi_impact_measures select * from SOURCE.indi_impact_measures orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_impact_measures COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_interdisciplinarity select * from SOURCE.indi_pub_interdisciplinarity orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_interdisciplinarity COMPUTE STATISTICS;
DROP TABLE IF EXISTS TARGET.result_new;

150
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/scripts/updateMonitorDBAll.sql
View File

@ -1,150 +0,0 @@
DROP TABLE IF EXISTS TARGET.result_new;
create table TARGET.result_new as
select distinct * from (
select * from SOURCE.result r where exists (select 1 from SOURCE.result_organization ro where ro.id=r.id and ro.organization in (
'openorgs____::4d4051b56708688235252f1d8fddb8c1', --Iscte - Instituto Universitário de Lisboa
'openorgs____::ab4ac74c35fa5dada770cf08e5110fab' -- Universidade Católica Portuguesa
) )) foo;
INSERT INTO TARGET.result select * from TARGET.result_new;
ANALYZE TABLE TARGET.result_new COMPUTE STATISTICS;
INSERT INTO TARGET.result select * from TARGET.result_new;
ANALYZE TABLE TARGET.result COMPUTE STATISTICS;
INSERT INTO TARGET.result_citations select * from SOURCE.result_citations orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_citations COMPUTE STATISTICS;
INSERT INTO TARGET.result_references_oc select * from SOURCE.result_references_oc orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_references_oc COMPUTE STATISTICS;
INSERT INTO TARGET.result_classifications select * from SOURCE.result_classifications orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_classifications COMPUTE STATISTICS;
INSERT INTO TARGET.result_apc select * from SOURCE.result_apc orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_apc COMPUTE STATISTICS;
INSERT INTO TARGET.result_concepts select * from SOURCE.result_concepts orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_concepts COMPUTE STATISTICS;
INSERT INTO TARGET.result_datasources select * from SOURCE.result_datasources orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_datasources COMPUTE STATISTICS;
INSERT INTO TARGET.result_fundercount select * from SOURCE.result_fundercount orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_fundercount COMPUTE STATISTICS;
INSERT INTO TARGET.result_gold select * from SOURCE.result_gold orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_gold COMPUTE STATISTICS;
INSERT INTO TARGET.result_greenoa select * from SOURCE.result_greenoa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_greenoa COMPUTE STATISTICS;
INSERT INTO TARGET.result_languages select * from SOURCE.result_languages orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_languages COMPUTE STATISTICS;
INSERT INTO TARGET.result_licenses select * from SOURCE.result_licenses orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_licenses COMPUTE STATISTICS;
INSERT INTO TARGET.result_oids select * from SOURCE.result_oids orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_oids COMPUTE STATISTICS;
INSERT INTO TARGET.result_organization select * from SOURCE.result_organization orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_organization COMPUTE STATISTICS;
INSERT INTO TARGET.result_peerreviewed select * from SOURCE.result_peerreviewed orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_peerreviewed COMPUTE STATISTICS;
INSERT INTO TARGET.result_pids select * from SOURCE.result_pids orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_pids COMPUTE STATISTICS;
INSERT INTO TARGET.result_projectcount select * from SOURCE.result_projectcount orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_projectcount COMPUTE STATISTICS;
INSERT INTO TARGET.result_projects select * from SOURCE.result_projects orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_projects COMPUTE STATISTICS;
INSERT INTO TARGET.result_refereed select * from SOURCE.result_refereed orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_refereed COMPUTE STATISTICS;
INSERT INTO TARGET.result_sources select * from SOURCE.result_sources orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_sources COMPUTE STATISTICS;
INSERT INTO TARGET.result_topics select * from SOURCE.result_topics orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_topics COMPUTE STATISTICS;
INSERT INTO TARGET.result_fos select * from SOURCE.result_fos orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_fos COMPUTE STATISTICS;
INSERT INTO TARGET.result_accessroute select * from SOURCE.result_accessroute orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.result_accessroute COMPUTE STATISTICS;
create or replace view TARGET.foo1 as select * from SOURCE.result_result rr where rr.source in (select id from TARGET.result_new);
create or replace view TARGET.foo2 as select * from SOURCE.result_result rr where rr.target in (select id from TARGET.result_new);
insert into TARGET.result_result select distinct * from (select * from TARGET.foo1 union all select * from TARGET.foo2) foufou;
drop view TARGET.foo1;
drop view TARGET.foo2;
ANALYZE TABLE TARGET.result_result COMPUTE STATISTICS;
-- indicators
-- Sprint 1 ----
INSERT INTO TARGET.indi_pub_green_oa select * from SOURCE.indi_pub_green_oa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_green_oa COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_grey_lit select * from SOURCE.indi_pub_grey_lit orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_grey_lit COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_doi_from_crossref select * from SOURCE.indi_pub_doi_from_crossref orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_doi_from_crossref COMPUTE STATISTICS;
-- Sprint 2 ----
INSERT INTO TARGET.indi_result_has_cc_licence select * from SOURCE.indi_result_has_cc_licence orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_has_cc_licence COMPUTE STATISTICS;
INSERT INTO TARGET.indi_result_has_cc_licence_url select * from SOURCE.indi_result_has_cc_licence_url orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_has_cc_licence_url COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_has_abstract select * from SOURCE.indi_pub_has_abstract orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_has_abstract COMPUTE STATISTICS;
INSERT INTO TARGET.indi_result_with_orcid select * from SOURCE.indi_result_with_orcid orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_with_orcid COMPUTE STATISTICS;
---- Sprint 3 ----
INSERT INTO TARGET.indi_funded_result_with_fundref select * from SOURCE.indi_funded_result_with_fundref orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_funded_result_with_fundref COMPUTE STATISTICS;
---- Sprint 4 ----
INSERT INTO TARGET.indi_pub_diamond select * from SOURCE.indi_pub_diamond orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_diamond COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_in_transformative select * from SOURCE.indi_pub_in_transformative orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_in_transformative COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_closed_other_open select * from SOURCE.indi_pub_closed_other_open orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_closed_other_open COMPUTE STATISTICS;
---- Sprint 5 ----
INSERT INTO TARGET.indi_result_no_of_copies select * from SOURCE.indi_result_no_of_copies orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_no_of_copies COMPUTE STATISTICS;
---- Sprint 6 ----
INSERT INTO TARGET.indi_pub_hybrid_oa_with_cc select * from SOURCE.indi_pub_hybrid_oa_with_cc orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_hybrid_oa_with_cc COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_bronze_oa select * from SOURCE.indi_pub_bronze_oa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_bronze_oa COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads select * from SOURCE.indi_pub_downloads orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads_datasource select * from SOURCE.indi_pub_downloads_datasource orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads_datasource COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads_year select * from SOURCE.indi_pub_downloads_year orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads_year COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_downloads_datasource_year select * from SOURCE.indi_pub_downloads_datasource_year orig where exists (select 1 from TARGET.result_new r where r.id=orig.result_id);
ANALYZE TABLE TARGET.indi_pub_downloads_datasource_year COMPUTE STATISTICS;
---- Sprint 7 ----
INSERT INTO TARGET.indi_pub_gold_oa select * from SOURCE.indi_pub_gold_oa orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_gold_oa COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_hybrid select * from SOURCE.indi_pub_hybrid orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_hybrid COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_has_preprint select * from SOURCE.indi_pub_has_preprint orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_has_preprint COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_in_subscribed select * from SOURCE.indi_pub_in_subscribed orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_in_subscribed COMPUTE STATISTICS;
INSERT INTO TARGET.indi_result_with_pid select * from SOURCE.indi_result_with_pid orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_result_with_pid COMPUTE STATISTICS;
INSERT INTO TARGET.indi_impact_measures select * from SOURCE.indi_impact_measures orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_impact_measures COMPUTE STATISTICS;
INSERT INTO TARGET.indi_pub_interdisciplinarity select * from SOURCE.indi_pub_interdisciplinarity orig where exists (select 1 from TARGET.result_new r where r.id=orig.id);
ANALYZE TABLE TARGET.indi_pub_interdisciplinarity COMPUTE STATISTICS;
DROP TABLE IF EXISTS TARGET.result_new;

12
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/scripts/updateMonitorDB_institutions.sql
View File

@ -1,12 +0,0 @@
DROP TABLE IF EXISTS TARGET.result_new;
create table TARGET.result_new as
select distinct * from (
select * from SOURCE.result r where exists (select 1 from SOURCE.result_organization ro where ro.id=r.id and ro.organization in (
'openorgs____::4d4051b56708688235252f1d8fddb8c1', --Iscte - Instituto Universitário de Lisboa
'openorgs____::ab4ac74c35fa5dada770cf08e5110fab' -- Universidade Católica Portuguesa
) )) foo;
INSERT INTO TARGET.result select * from TARGET.result_new;
ANALYZE TABLE TARGET.result_new COMPUTE STATISTICS;

110
dhp-workflows/dhp-stats-monitor-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats-monitor/oozie_app/workflow.xml
View File

@ -1,110 +0,0 @@
<workflow-app name="Stats Monitor Update" xmlns="uri:oozie:workflow:0.5">
<parameters>
<property>
<name>stats_db_name</name>
<description>the target stats database name</description>
</property>
<property>
<name>monitor_db_name</name>
<description>the target monitor db name</description>
</property>
<property>
<name>monitor_db_shadow_name</name>
<description>the name of the shadow monitor db</description>
</property>
<property>
<name>hive_metastore_uris</name>
<description>hive server metastore URIs</description>
</property>
<property>
<name>hive_jdbc_url</name>
<description>hive server jdbc url</description>
</property>
<property>
<name>hive_timeout</name>
<description>the time period, in seconds, after which Hive fails a transaction if a Hive client has not sent a hearbeat. The default value is 300 seconds.</description>
</property>
<property>
<name>hadoop_user_name</name>
<description>user name of the wf owner</description>
</property>
</parameters>
<global>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<configuration>
<property>
<name>hive.metastore.uris</name>
<value>${hive_metastore_uris}</value>
</property>
<property>
<name>hive.txn.timeout</name>
<value>${hive_timeout}</value>
</property>
<property>
<name>mapred.job.queue.name</name>
<value>analytics</value>
</property>
</configuration>
</global>
<start to="resume_from"/>
<decision name="resume_from">
<switch>
<case to="Step1-updateMonitorDB">${wf:conf('resumeFrom') eq 'Step1-updateMonitorDB'}</case>
<case to="Step2-copyDataToImpalaCluster">${wf:conf('resumeFrom') eq 'Step2-copyDataToImpalaCluster'}</case>
<case to="Step3-finalizeImpalaCluster">${wf:conf('resumeFrom') eq 'Step3-finalizeImpalaCluster'}</case>
<default to="Step1-updateMonitorDB"/>
</switch>
</decision>
<kill name="Kill">
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="Step1-updateMonitorDB">
<shell xmlns="uri:oozie:shell-action:0.1">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<exec>monitor.sh</exec>
<argument>${stats_db_name}</argument>
<argument>${monitor_db_name}</argument>
<argument>${monitor_db_shadow_name}</argument>
<argument>${wf:appPath()}/scripts/updateMonitorDB_institutions.sql</argument>
<argument>${wf:appPath()}/scripts/updateMonitorDB.sql</argument>
<argument>${wf:appPath()}/scripts/updateMonitorDBAll.sql</argument>
<file>monitor.sh</file>
</shell>
<ok to="Step2-copyDataToImpalaCluster"/>
<error to="Kill"/>
</action>
<action name="Step2-copyDataToImpalaCluster">
<shell xmlns="uri:oozie:shell-action:0.1">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<exec>copyDataToImpalaCluster.sh</exec>
<argument>${monitor_db_name}</argument>
<argument>${hadoop_user_name}</argument>
<file>copyDataToImpalaCluster.sh</file>
</shell>
<ok to="Step3-finalizeImpalaCluster"/>
<error to="Kill"/>
</action>
<action name="Step3-finalizeImpalaCluster">
<shell xmlns="uri:oozie:shell-action:0.1">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<exec>finalizeImpalaCluster.sh</exec>
<argument>${monitor_db_name}</argument>
<argument>${monitor_db_shadow_name}</argument>
<file>finalizeImpalaCluster.sh</file>
</shell>
<ok to="End"/>
<error to="Kill"/>
</action>
<end name="End"/>
</workflow-app>