sabeel
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

entity resolution: added distcpt of missing entities in graph materialization

This commit is contained in:
Sandro La Bruzzo 2021-11-22 15:55:24 +01:00
parent 93fe8ce8b2
commit 483d3039d1
1 changed files with 36 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/resolution/oozie_app/workflow.xml
View File

@ -69,10 +69,44 @@
<arg>--workingPath</arg><arg>${workingDir}</arg> <arg>--workingPath</arg><arg>${workingDir}</arg>
<arg>--targetPath</arg><arg>${targetPath}</arg> <arg>--targetPath</arg><arg>${targetPath}</arg>
</spark> </spark>
<ok to="End"/> <ok to="copy_entities"/>
<error to="Kill"/> <error to="Kill"/>
</action> </action>
<end name="End"/> <fork name="copy_entities">
<path start="copy_organization"/>
<path start="copy_projects"/>
<path start="copy_datasources"/>
</fork>
<action name="copy_organization">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<arg>${nameNode}/${graphBasePath}/organization</arg>
<arg>${nameNode}/${targetPath}/organization</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<action name="copy_projects">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<arg>${nameNode}/${graphBasePath}/project</arg>
<arg>${nameNode}/${targetPath}/project</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<action name="copy_datasources">
<distcp xmlns="uri:oozie:distcp-action:0.2">
<arg>${nameNode}/${graphBasePath}/datasource</arg>
<arg>${nameNode}/${targetPath}/datasource</arg>
</distcp>
<ok to="copy_wait"/>
<error to="Kill"/>
</action>
<join name="copy_wait" to="End"/>
<end name="End"/>
</workflow-app> </workflow-app>