diff --git a/dump/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/eoscdump/oozie_app/workflow.xml b/dump/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/eoscdump/oozie_app/workflow.xml index f0e4970..c133ff9 100644 --- a/dump/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/eoscdump/oozie_app/workflow.xml +++ b/dump/src/main/resources/eu/dnetlib/dhp/oa/graph/dump/eoscdump/oozie_app/workflow.xml @@ -153,13 +153,38 @@ --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} --sourcePath${sourcePath} - --resultTableNameeu.dnetlib.dhp.schema.oaf.Publication --resultPath${workingDir}/dump/publication + --outputPath${workingDir}/dump/publicationextendedaffiliation + + + + + + + yarn + cluster + Extend Dump Publication with indicators + eu.dnetlib.dhp.oa.graph.dump.eosc.ExtendWithUsageCounts + dump-${projectVersion}.jar + + --executor-memory=${sparkExecutorMemory} + --executor-cores=${sparkExecutorCores} + --driver-memory=${sparkDriverMemory} + --conf spark.extraListeners=${spark2ExtraListeners} + --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} + --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} + --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} + --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} + + --actionSetPath${actionSetPath} + --resultPath${workingDir}/dump/publicationextendedaffiliation --outputPath${workingDir}/dump/publicationextended + + yarn @@ -203,13 +228,38 @@ --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} --sourcePath${sourcePath} - --resultTableNameeu.dnetlib.dhp.schema.oaf.Dataset --resultPath${workingDir}/dump/dataset + --outputPath${workingDir}/dump/datasetextendedaffiliation + + + + + + + yarn + cluster + Extend Dump Dataset with indicators + eu.dnetlib.dhp.oa.graph.dump.eosc.ExtendWithUsageCounts + dump-${projectVersion}.jar + + --executor-memory=${sparkExecutorMemory} + --executor-cores=${sparkExecutorCores} + --driver-memory=${sparkDriverMemory} + --conf spark.extraListeners=${spark2ExtraListeners} + --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} + --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} + --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} + --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} + + --actionSetPath${actionSetPath} + --resultPath${workingDir}/dump/datasetextendedaffiliation --outputPath${workingDir}/dump/datasetextended + + yarn @@ -254,12 +304,37 @@ --sourcePath${sourcePath} --resultPath${workingDir}/dump/otherresearchproduct - --resultTableNameeu.dnetlib.dhp.schema.oaf.OtherResearchProduct + --outputPath${workingDir}/dump/otherresearchproductextendedaffiliation + + + + + + + yarn + cluster + Extend Dump ORP with indicators + eu.dnetlib.dhp.oa.graph.dump.eosc.ExtendWithUsageCounts + dump-${projectVersion}.jar + + --executor-memory=${sparkExecutorMemory} + --executor-cores=${sparkExecutorCores} + --driver-memory=${sparkDriverMemory} + --conf spark.extraListeners=${spark2ExtraListeners} + --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} + --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} + --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} + --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} + + --actionSetPath${actionSetPath} + --resultPath${workingDir}/dump/otherresearchproductextendedaffiliation --outputPath${workingDir}/dump/otherresearchproductextended + + yarn @@ -304,12 +379,37 @@ --sourcePath${sourcePath} --resultPath${workingDir}/dump/software - --resultTableNameeu.dnetlib.dhp.schema.oaf.Software + --outputPath${workingDir}/dump/softwareextendedaffiliation + + + + + + + yarn + cluster + Extend Dump ORP with indicators + eu.dnetlib.dhp.oa.graph.dump.eosc.ExtendWithUsageCounts + dump-${projectVersion}.jar + + --executor-memory=${sparkExecutorMemory} + --executor-cores=${sparkExecutorCores} + --driver-memory=${sparkDriverMemory} + --conf spark.extraListeners=${spark2ExtraListeners} + --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} + --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} + --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} + --conf spark.sql.warehouse.dir=${sparkSqlWarehouseDir} + + --actionSetPath${actionSetPath} + --resultPath${workingDir}/dump/softwareextendedaffiliation --outputPath${workingDir}/dump/softwareextended + +