D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

[graph2hive] more partitions

This commit is contained in:
Claudio Atzori 2022-10-14 15:47:58 +02:00
parent b47aaf4dd1
commit ae7cd0735a
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/hive/oozie_app/workflow.xml
View File

@ -126,6 +126,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.Publication</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>8000</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>
@ -152,6 +153,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.Dataset</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>4000</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>
@ -178,6 +180,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.OtherResearchProduct</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>3000</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>
@ -204,6 +207,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.Software</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>300</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>
@ -230,6 +234,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.Datasource</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>100</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>
@ -256,6 +261,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.Organization</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>400</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>
@ -309,6 +315,7 @@
<arg>--hiveDbName</arg><arg>${hiveDbName}</arg>
<arg>--className</arg><arg>eu.dnetlib.dhp.schema.oaf.Relation</arg>
<arg>--hiveMetastoreUris</arg><arg>${hiveMetastoreUris}</arg>
<arg>--numPartitions</arg><arg>10000</arg>
</spark>
<ok to="join_import"/>
<error to="Kill"/>