dimitris.pierrakos
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

fixed input path supplemented to GetDatasourceFromCountry; adjusted the various spark.sql.shuffle.partitions

This commit is contained in:
Claudio Atzori 2023-03-24 13:09:12 +01:00
parent c07857fa37
commit b502f86523
1 changed files with 10 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/clean/oozie_app/workflow.xml
View File

@ -109,9 +109,9 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=10000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphOutputPath}</arg> <arg>--inputPath</arg><arg>${graphInputPath}</arg>
<arg>--workingDir</arg><arg>${workingDir}/working/hostedby</arg> <arg>--workingDir</arg><arg>${workingDir}/working/hostedby</arg>
<arg>--country</arg><arg>${country}</arg> <arg>--country</arg><arg>${country}</arg>
</spark> </spark>
@ -160,7 +160,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=15000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/publication</arg> <arg>--inputPath</arg><arg>${graphInputPath}/publication</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/publication</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/publication</arg>
@ -194,7 +194,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=8000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/dataset</arg> <arg>--inputPath</arg><arg>${graphInputPath}/dataset</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/dataset</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/dataset</arg>
@ -228,7 +228,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=5000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/otherresearchproduct</arg> <arg>--inputPath</arg><arg>${graphInputPath}/otherresearchproduct</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/otherresearchproduct</arg>
@ -262,7 +262,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=2000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/software</arg> <arg>--inputPath</arg><arg>${graphInputPath}/software</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/software</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/software</arg>
@ -296,7 +296,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=1000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/datasource</arg> <arg>--inputPath</arg><arg>${graphInputPath}/datasource</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/datasource</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/datasource</arg>
@ -330,7 +330,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=1000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/organization</arg> <arg>--inputPath</arg><arg>${graphInputPath}/organization</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/organization</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/organization</arg>
@ -364,7 +364,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=2000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/project</arg> <arg>--inputPath</arg><arg>${graphInputPath}/project</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/project</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/project</arg>
@ -398,7 +398,7 @@
--conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners} --conf spark.sql.queryExecutionListeners=${spark2SqlQueryExecutionListeners}
--conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress} --conf spark.yarn.historyServer.address=${spark2YarnHistoryServerAddress}
--conf spark.eventLog.dir=${nameNode}${spark2EventLogDir} --conf spark.eventLog.dir=${nameNode}${spark2EventLogDir}
--conf spark.sql.shuffle.partitions=7680 --conf spark.sql.shuffle.partitions=20000
</spark-opts> </spark-opts>
<arg>--inputPath</arg><arg>${graphInputPath}/relation</arg> <arg>--inputPath</arg><arg>${graphInputPath}/relation</arg>
<arg>--outputPath</arg><arg>${graphOutputPath}/relation</arg> <arg>--outputPath</arg><arg>${graphOutputPath}/relation</arg>